This study evaluates the spatial–temporal differences of carbon emissions and ecosystem survives value (ESV) induced by land use cover change in Chengyu urban agglomeration of China, during which uncertainties expressed as interval rough sets are integrated into the modeling framework. Results reveal that built-up land accounts for [88.62%, 99.64%] of total carbon emissions. Compared with other cities, the higher amount of carbon emissions occur in Chengdu and Chongqing, while the eastern region has a stronger carbon absorption capacity. Uncertainties in carbon-emission coefficient have significant effects on regional carbon emissions. With aid of interval-stochastic analysis for carbon emersions, it can help offset some of the mutual effects caused by differences between cities. The values of eco-carrying and carbon-emission coefficients are characterized as a feature of high to low from the east to west, and their highest values exist in Liangshan Yi Autonomous Prefecture. There are slight variations in the regional average ESVs during the periods from 2000 to 2015, but a high sensitive coefficient occurs in grassland that has a significant influence on the overall ESV. Generally, a high regional economic performance corresponds to a high amount of carbon emissions and a low ESV, but variations in the proportion of cultivated and forest lands have slight impacts on carbon emissions and ecosystem services. Policy implications indicate that switching from coal to shale gas might be regarded as a potential approach for carbon-emission reduction. Regional cooperation also is critical channel to promote regional green development of Chengyu urban agglomeration.