Abstract: “Regular cluster regions” of data-point sets containing both order attributes and out-of-order attributes are obtained by applying decision tree method.Feature weight of both order attributes and out-of-order attributes can be optimized separately by using the principia of “data points in any cluster are close to each other，and data points between any two clusters are away from each other”.This optimization method of feature weight based on decision tree division can optimize feature weight of data sets with hybrid attributes.