Based on panel data of China's 31 provinces from 2007 to 2016, this study uses VRS model to assess provincial agricultural science and technology innovation performance. Then it analyzes spatial and temporal disparity of technology innovation performance in six regions, namely, the Eastern, the Northwest, the Northern, the South Central, the Southwest and the Northeast, and conducts σ convergence and β convergence analysis of technology innovation performance differences in research institutions. The empirical results are as follows:In terms of spatial differences, there are significant differences between provincial performances. According to its performance value from high to low, six regions can be sorted as follows, the Eastern > the Northwest > the Northern > the South Central > the South West > the Northeast. It didn't show significant σ convergence in overall disparity of science and technology innovation performance among regional agricultural research institutions. But absolute β convergence and conditional β convergence model suggest that difference of science and technology innovation performance of agricultural research institutions is gradually narrowing down between six regions. And there exists "chase effect" of backward areas on advanced regions. Provinces in North China will narrow the gap between each other at the fastest speed to achieve internal homeostasis level, followed by the Northeastern. And that of the Eastern region is the slowest. A proper understanding of spatial and temporal differences of science and technology innovation performance in China agricultural scientific research institutions is an important prerequisite for rationally allocating agricultural science and technology resources, strengthening the driving force of agricultural science and technology innovation, improving the innovation system of agricultural science and technology, speeding up the pace of modern agriculture, and promoting the strategy of Rural Revitalization.