Random forest is a classifier structure, which can use tree nodes to organize the training and prediction processing of sample data. The advantages of random forest are to overcome the disadvantages of complex classification rules, convergence to non-global local optimal solution and overfitting in decision trees, and combine single classifiers into multiple classifiers. Random forest algorithm is an effective extension of decision tree algorithm, and the integration of ID3 and CART algorithm effectively improves the robustness of model classification and prediction ability. This paper proposes a novel random forest algorithm which combines ID3 algorithm and CART algorithm. Finally, the public test data set is selected to carry out experiments. The experimental results show that the new random forest algorithm proposed in this paper has a significant improvement in accuracy, precision and recall rate.