Based on the panel data of 35 important cities in China from 2003 to 2015, taking agricultural added value, industrial added value and service industry added value excluding real estate as the explained variables, and taking the degree of land financial dependence as the explanatory variable, this paper constructs a panel model across cities to effectively explore the industrial depression effect of land finance, and tests the stability, endogeneity and regional heterogeneity of empirical results. The empirical study finds that：（1）The degree of land finance dependence has a significant negative effect on the agricultural added value, industrial added value and service industry added value excluding real estate, and the estimation result has passed the robustness test, which shows that land finance has significant industrial depression effect.（2）Using the instrumental variable estimation method to test endogeneity, it is found that the industrial depression effect of land finance still exists significantly, and after the endogenous factors are overcome, the industrial depression effects are all obviously strengthened. In the investigation of regional heterogeneity, it is found that there is regional heterogeneity in the industrial depression effects, and the regional heterogeneity mainly shows that the western region has the obvious depression effect only in the service industry, and there is the significant promotion effect in agriculture and industry.（3）When analyzing the control variables, whether it is robustness, endogenous or regional heterogeneity test, the wage level and the logistic development level have positive effect on the industrial development. the positive effect of government size on agriculture and industry in the western region exists, it is estimated that excessive government size is not conducive to the rapid development of the three major industries.