Based on the resource mismatch theory, a mismatch measurement model is constructed to measure the mismatch index of agricultural science and technology resources, and ArcGIS software and Markov chain analysis method are used to analyze the space-time evolution of the mismatch. It is found that：① From 2005 to 2019, the space-time evolution of human resource mismatch in each province and the three major regions showed a divergence between excess and deficit, while the overall evolutionary pattern is that excess mismatch dominates and gradually worsens; the financial resource mismatch as a whole is in an evolutionary feature where under-allocation and excess are comparable and relatively stable.② Markov chain analysis found that both re?sources had a tendency to evolve in space-time in a way that maintains the stability of the original mismatch hierarchy state, where the stability of human mismatch is stronger than that of financial resources, and the stability of moderate and severe mismatch is very strong. There is a widespread mismatch of agricultural science and technology resources between regions and between provinces within regions, and the evolution of the mismatch is varies, and it is difficult to show significant improvement. To this end, it is suggested to multi-measures to accelerate the free flow of human resources in agricultural science and technology across regions and provinces and optimize the allocation of financial resources in agricultural science and technology,in order to alleviate the mismatch of agricultural science and technology resources between regions and provinces within regions, and thus improve agricultural science and technology output.