Abstract: The flat-plate solar collector useful energy gain shows a certain degree of intermittence,randomness and periodicity due to the influence of external factors such as weather.The method of probabilistic model can reflect above the characteristics more accurately.However,the parameter analysis method in current probability models should be assumed the prior probability distribution,which is not universality.For this reason,the probabilistic flat-plate solar collector useful energy gain model based on nonparametric kernel density estimation is proposed.The golden section algorithm is used to solve the optimal bandwidth of nonparametric kernel density estimation.The correctness and reliability of the model are verified by the goodness of fit test and error test.Solar collector useful energy gain sample data can be got from the dynamic simulation model platform of TRNSYS,which verifies the effectiveness of the optimal bandwidth of the probabilistic flat-plate solar collector useful energy gain model based on nonparametric kernel density estimation and the adaptability of the model at different time scales.