Early diagnosis of breast cancer has been playing very important role on treatment of the disease. In this work, a new feature selection method for breast cancer classification based on data fusion and genetic algorithm is presented. The study consists of two steps: In the first step, the dimensionality of the gene expression dataset was reduced with filter method and the second step, significant genes have been identified with genetic algorithm. SVM was used for fitness function in genetic programming. In this study the classification accuracy rate was obtained 94.65 % when using selected 10 genes.