As a hot research topic in the field of pattern recognition, feature selection is an important method to dimension reduction. The current research is mainly focused on the discrete features. In this paper, a feature selection method for continuous features is proposed by introducing the concept of similarity degrees of interval numbers. Based on the similarity degrees of interval numbers, this method redefines each feature's attribute similarity as heuristic information of feature selection. Then, achieve the goal by ranking the feature corpus to select the feature subset. The experiments on the UCI Repository data sets have demonstrated that the approach of the feature ranking and feature selection has greatly improved the effectiveness and efficiency of classifications on continuous features.