On the basis of empirical data concerning "four agricultural subsidies" in Chongqing City, we establish panel data model to analyze the general effect of "four agricultural subsidies" on grain yield, and the structural effect of "four agricultural subsidies" on grain yield. Based on the F-test and Hausman test, we estimate the model using GLS or FGLS method. The results show that "four agricultural subsidies" have significant positive effect on grain yield; the structural effect of "four agricultural subsidies" on grain yield is different (comprehensive subsidies for means of agricultural production has the strongest effect on grain yield, followed by subsidies for promotion of improved variety of seed; the effect of subsidies for purchase of agricultural machinery on grain yield is not significant). Therefore, governments at all levels should continue to strengthen the support efforts in "four agricultural subsidies", and continually optimize the structure to increase grain yield.