Abstract: The concept of attribute similarity is presented to solve the weight problem of multi-dimension object classification, and cloud classifier based on attributes similarity is presented also. Every attribute model of training set is set up by cloud model, which describes the membership to which any an attribute value belongs its class center Ex. Classification model is integrated by every attribute model, and attribute weight is calculated by attribute similarity. The larger the similarity of an attribute, the smaller the action of the attribute to classification is. The center of classifier is optimized by particle swarm optimization algorithm. The classifier is used in classifying iris data set, its classification result is better than that of the common cloud model.