Abstract: This paper applied input-output method,spatial autocorrelation method and path analysis model to make a study of China's provincial spatial and temporal evolution characteristics of carbon lock-in and its effect mechanism based on carbon inputs and carbon output perspectives. The results showed that each China's province's carbon lock-in situation had been significantly improved but unbalanced. The carbon lock-in situation in eastern areas is relatively high,central region is less high,and west regions is low,and the gap between the eastern region and the Midwest region has widened. The carbon lock-in situation has a significant spatial correlation characteristic. The spatial evolution pattern became more and more centralized from disperse and its Matthew effect has been weakened. The improvement of innovation,industrial structure,technological progress and energy efficiency has a positive role on promoting the carbon unlocking effect. Energy efficiency,technological progress and industrial structure generate a greater carbon unlocking effect directly. The innovation level,technological progress and industrial structure generate an obvious carbon unlocking effect indirectly.