This study selected Picea crassifolia forest which in the Pailugou catchment of the middle of the northern slope of the Qilian Mountains as objects. Based on the observation data of rainfall interception, meteorological data and forestland investigation data in the growing season (May to September) of 2019, the rainfall interception process was analyzed, and the rainfall interception process of P. crassifolia forest was simulated by using modified Gash model. The results showed that the total amount of atmospheric rainfall outside the forest during the study period was 418.7 mm, and the rainfall was mainly moderate and light rain. The throughfall, trunk stem flow and canopy interception was 284.5, 6.7 and 127.5 mm, respectively, accounting for 67.94%, 1.60% and 30.46% of the total atmospheric rainfall, respectively. According to the simulation results of modified Gash model, the corresponding simulated value of throughfall, trunk stem flow and canopy interception was 302.2, 3.7 and 112.8 mm, respectively, accounting for 72.18%, 0.88% and 26.94% of the total atmospheric rainfall, respectively. Compared with the measured values, the relative errors of the above parameters were 6.22%, 44.78% and 11.53%. According to the parameters sensitivity analysis of the modified Gash model, the influence of the parameters on the canopy interception rate during the change process of -50% to 50% followed the order of E > c > R > S > S_t > P_t. In conclusion, the modified Gash model showed a good fitting effect in the simulation of rainfall interception of P. crassifolia forest in Qilian Mountains. The research results provide references for the research of the water cycle of P. crassifolia forest in Qilian Mountains.