Abstract: Sensitivity analysis, which can quantitatively estimate the contribution of input variable to the output, is an effective way to reveal the inherent laws of the model. In this paper, sensitivity analysis was applied to the algal blooms forecast model based on BP neural network in Chaohu Lake. The result of the analysis indicates that algal blooms in Chaohu are affected by many factors. There was a positive correlation among the change of water temperature, dissolved oxygen, turbidity, atmospheric temperature, illumination and the change of mass concentration of algal. Among these factors, atmospheric temperature is the most important, with a relative contribution up to 17.01%; on the contrary, the rise of atmospheric pressure does harm to the algal; and the influence of high pH on the algal concentration is uncertain.