Abstract: Based on the Cobb-Douglas production function model, this paper introduces the public infrastructure capital stock into the economic growth model, uses the spatial panel data model effect analysis method, and uses the statistics of the Yangtze River Economic Belt from 2003 to 2016 to examine different types of public. The relationship between infrastructure capital stock and regional economic growth in different periods. The empirical results show that: (1) The economic growth of the provinces and cities in the Yangtze River Economic Belt has significant spatial dependence characteristics. After 2008, the degree of spatial dependence and its significance gradually declined. (2) Different types of public infrastructure capital stocks have significant differences in the spatial effects of regional economic growth. The energy infrastructure capital stock significantly promoted the global economic growth with a total effect of 0.515; the transportation infrastructure capital stock significantly boosted the local economic growth, the direct effect was 0.067 8, the indirect effect was not obvious; the water environment infrastructure capital stock inhibited the local economic growth, Pulling the economic growth in the adjacent region, the direct effect is -0.047 3, and the indirect effect is 0.165. (3) In different periods, with the accumulation of public infrastructure capital stock, its spatial effect on economic growth has changed significantly. In the period 2003-2008 and the period of 2009-2016, the spatial effect of the capital stock of energy infrastructure on regional economic growth has been turned negative, indicating that its stock has exceeded the optimal investment scale, and there is a phenomenon of scale uneconomic; The comparison of the capital stock of transportation and water environment infrastructure in two time periods shows that its spatial effect on economic growth is gradually increasing with the increase of investment volume.