Abstract: To predict the amount of dust accumulated on the surface of photovoltaic modules,the meteorological factors were used as variables,the Least Squares Support Vector Machine（LSSVM） was used as Prediction algorithm,and by improving the attenuation method of the inertia factor in the particle swarm optimization（PSO） to improve the algorithm’s optimization robustness.In the prediction model of dust accumulation,local time and space factors were taken account,combined the rainfall which was strongly correlated and the power attenuation rate of photovoltaic modules to establish a dust accumulation prediction model under natural rainfall.Measurement experiments were carried out in Hangzhou,and the example calculation showed that the model could quickly predict the amount of dust accumulation of photovoltaic modules and the rate of power attenuation caused by dust accumulation during the statistical period,which could provide a basis for accurately predicting the power generation of photovoltaic modules and formulating the cleaning frequency of modules.