Objective As the core spatial unit responsible for supporting fundamental urban operations, urban areas accommodate a wide range of human activities—residential, industrial, commercial, and transportation—each associated with specific land use types. The functional operation of these different land use types dominates the majority of urban energy consumption and carbon emissions, making them the primary sources of urban carbon output and critical units for systematic low-carbon land use planning. Achieving the national “dual-carbon” strategic goals (“carbon peak and carbon neutrality”) requires precise carbon emission management and effective planning interventions at the urban area scale. However, existing research on carbon emissions accounting and driving mechanisms remains largely concentrated at the national, provincial, or municipal levels, with limited focus on the urban scale and insufficient linkage to specific land use types. This gap hinders the formulation of targeted, spatially explicit low-carbon land use planning strategies. Therefore, it is of critical importance to conduct fine-grained accounting of carbon emissions from different land use types within urban areas, analyze their spatiotemporal dynamics and evolutionary characteristics, and identify the underlying driving mechanisms. These works provide a scientific basis and decision support for achieving low-carbon urban development and refining land use planning strategies.

Methods This study selects the Cai Jia Smart New City—a typical and representative urban unit within Chongqing—as the research object. We constructed an urban-scale carbon emissions accounting model that integrates “Human Activity−Land Use Type−Carbon Emissions.” This framework systematically links socioeconomic activities to their corresponding land use categories, enabling accurate attribution of carbon emissions. Based on this model, we calculated the actual carbon emissions for four key land use modules—residential, industrial, commercial service, and transportation—for the years 2016 and 2021. Furthermore, carbon emissions for the year 2035 were projected under the planned land use plan. The analysis compared changes across multiple dimensions: total carbon emissions driven by shifts in land use types, total emissions from different land use categories, carbon emission intensity (emissions per unit area), carbon intensity per unit of GDP, and per capita carbon emission intensity. To delve into the driving forces behind these changes, the logarithmic mean divisia index (LMDI) decomposition method was employed. This technique quantitatively analyzes the temporal evolution and contribution rates of key driving factors—including land use scale, economic development level, and population concentration—to changes in land use carbon emissions. Subsequently, guided by multi-objective planning principles, the LINGO mathematical optimization software was utilized to develop a land use structure optimization model. With the dual objective functions of economic benefit maximization and carbon emission minimization, and under a set of constraints reflecting local natural conditions and development policies, this study derived an optimized land use structure scheme for Chongqing’s Caijia Smart New City for the target year of 2035.

Results The analysis results show that 1) the carbon emission intensity of industrial land in Caijia Smart New City is the largest and shows a continuous downward trend, while other land use show an upward trend from 2016 to 2021 and a downward trend from 2021 to 2035; the total amount of carbon emissions from land use, carbon emissions per unit of gross domestic product (GDP), and the total amount of carbon emissions per capita show a downward trend; 2) per capita output value has a pulling effect on industrial land and commercial land; the energy efficiency of land use has a pulling effect on residential land; and the energy efficiency of land use has a pulling effect on residential land; 3) per capita output value has a pulling effect on industrial land and commercial land; land use energy efficiency has a pulling effect on residential land and transportation land; output value density has an inhibiting effect on commercial land and transportation land; energy structure has an inhibiting effect on residential land; land use energy efficiency has an inhibiting effect on industrial land.

Conclusion Based on the accounting and decomposition results, an optimized land use structure plan was constructed using the multi-objective optimization model. Compared to the 2035 land use planning plan, the optimized plan achieves a reduction in per capita carbon emission by 0.07 ton/person while maintaining the same level of carbon emission per unit of GDP. This outcome demonstrates that the optimized scheme can effectively lower carbon emissions while largely satisfying the requirements of economic development. Finally, an integrated triple strategy encompassing “industrial structure optimization, energy structure adjustment, and land use energy efficiency improvement” is proposed. These strategies are targeted and elaborated specifically for industrial, residential, commercial service, and transportation land modules. This study aims to enrich and improve the methodological system for urban-scale carbon emission accounting and low-carbon land use planning based on land use types. The findings and framework provide theoretical support and practical guidance for future low-carbon-oriented land use planning and policy formulation in urban areas.