In this paper, the theory and method of fuzzy mathematics are applied to forecast the activity of  solar active region. According to the correlation between the flares and the characteristic factors of the sunspot groups, the membership functions are constructed to comprehensively evaluate and predict the activity of solar active regions. The precise results were obtained by data reduction and analysis with the accuracy  higher than 95% after comparing with observations. An applicable software of the data reduction is successfully compiled for the solar activity prediction. Thus, when the values of characteristic factors of the sunspot groups are inputted into the computer, the predicting results of the activity of solar active regions are instantly obtained. Forecast results indicate that the method of the fuzzy comprehensive evaluation is a good one for the solar activity prediction.