Fluorescence spectrum of mixed solution between chlorothalonil and typical fruit juice (apple juice and peach juice) were obtained with fluorescence spectrophotometer. It was found that there was a characteristic peak in 352 nm of chlorothalonil. Regression analysis was applied in the modeling of relationship between fluorescence intensity and chlorothalonil concentration. Estimation function of chlorothalonil concentration was deduced through fluorescence spectrum and its derivative fluorescence spectrum. The correlation coefficients of the exponential prediction model under two kinds of spectral patterns were higher than 0.99, which was better than the linear function model. For the two kinds of fruit juice, the average recoveries of the exponential model function under the original spectral pattern were 101% and 100% while the average recoveries of the linear model were 110% and 118%, respectively; The average recoveries of the exponential model function under the derivative spectral mode were 101% and 102%, and the average recoveries of the linear model were 109% and 120% respectively. The analysis results showed that fluorescence spectrometry can be used to detect and predict the chlorothalonil residue in fruit juice, and the performance of established exponential function model was better than the linear function model. At the same time, derivative fluorescence spectrometry method was found to have no significant advantage in the concentration prediction model of chlorothalonil residue in fruit juice, so the original fluorescence spectrum can be directly applied in the modeling analysis.