Abstract: The modified general soil loss equation (RUSLE) is widely used in estimating regional soil erosion, but the soil and water conservation measure factor (P), as the most difficult factor to determine, is mainly assigned by literature results, land use type and slope, and there is still room for improvement in accuracy. In order to optimize the value of soil and water conservation measures factor (P) of RUSLE model and improve the simulation accuracy of the model. According to the definition of P, a series of P values were obtained from soil erosion monitoring data of runoff plots and control plots with different soil and water conservation measures at 8 monitoring stations in Huanggang, Wuhan, Shiyan and Xianning in Hubei Province, and DEM, soil type, land use, vegetation cover and terrace data sets were fused layer by layer to match the above obtained P values to the corresponding grid cells, and then optimize the value of P factors to obtain the spatial distribution map of P factors in Hubei Province. Then, based on the P value before optimization and the P value after optimization, the RUSLE model was used to calculate the soil loss in Hubei Province, and the soil erosion modulus before and after optimization was obtained. Finally, the P value was optimized based on the measured soil loss evalua-tion in runoff plots and small watersheds. RUSLE model simulation accuracy. It is feasible to use the measured data of runoff plot combined with multi-source data to optimize the value of P factor of RUSLE model. The proportions of high P values (≥ 0. 600) were 6.540% (0.6), 0.533% (0.7), 0.083% (0.8), and 0.002% (1), respectively, distributed in the west, northeast and southeast of Hubei Province, with lower P values. The areas are located in the central plain.; Before and after optimization, the temporal and spatial distribution pattern of soil erosion in Hubei Province is basically consistent with the key areas where soil erosion occurs. The accuracy evaluation based on the observation data of runoff plots shows that R2 is 0.816, RMSE is 214.6, and the relative error is 11.22%. 98% of the data points in the optimized RUSLE prediction results are located in the 95% prediction band; The accu-racy evaluation based on small watershed observation data shows that the optimized simulation accu-racy reaches 86.630%, which is 9.798% higher than that before optimization; The proportion of strong erosion, extremely strong erosion and severe erosion before optimization is greater than that after op-timization, because the P factor before optimization is higher than that after optimization. The results show that the optimized P-factor value method is effective, and the simulation accuracy of RUSLE model is improved.