To optimize the accuracy of computer aided diagnosis (CAD) technology in the diagnosis of breast cancer, we propose a new support vector machine algorithm   based on the feature weighting of Gini index under the random forest model (RFG-SVM). The algorithm uses the Gini index under the random forest model to measure the impact of each feature on the classification results, and to build a support vector machine with the weighted feature vector kernel function, which is then applied to the diagnosis of breast cancer. Theoretical analysis and experimental data tests show that the proposed algorithm has higher classification accuracy than the traditional SVM and is more competitive than the state-of-the-art methods in medical diagnostics.