In this paper, the theory and method of fuzzy clustering analysis are applied to study the properties of solar active regions. By use of the data of soft X-ray flares and the characteristic factors of the sunspot groups in the SGD, the standardized value is obtained by average value and standard deviation. The fuzzy similitude matrix and equivalent matrix are built by use of the included angle cosine transform and the method of similitude coefficient, then we can divide 24 active regions into 5 grades of activity according to λ threshold valuation. The activity prediction of active regions has been obtained by data reduction and analysis.