This study aimed at finding a common specific indication of tongue images between model rats and patients with Qi deficiency and blood stasis syndrome of CHD. Methods Images of tongues of patients with Qi deficiency and blood stasis syndrome of CHD were divided by mat lab software and relevant features were extracted. Whole clinical and animal samples were classified and identified using an RF model and the predictive value of the model was evaluated by the ROC. AUC was calculated to determine the diagnostic value of each index. Results Four indications named M1 , M2 , M3 and M5 were closely related to the diagnosis of CHD with Qi deficiency and blood stasis syndrome. In RF classification, FPR was 15. 2%, TPR was 83. 3%, the best boundary value was 0. 5, sensitivity was 81. 4%, specificity was 76%, accuracy was 85. 7%, circumferential convolution was 0. 6991, and AUC was 0. 8151. AUC indicated that all variables had diagnostic value, among which M1 , M2 and M5 had high diagnostic values and M3 had a moderate diagnostic value. In the rat model, significant differences in the characteristic values of M1 , M2 , M3 and M5 were observed compared with the normal group (P < 0. 05, P < 0. 01). In RF classification, FPR was 14. 8%, TPR was 83. 1%, AUC was 0. 8755, the best boundary value was 0. 5, sensitivity was 81. 8%, specificity was 75. 6%, accuracy was 83. 1%, and circumferential convolution was 0. 6451. AUC suggested that M1 , M2 , M3 and M5 had diagnostic value. Among them, M1 , M2 and M5 had high diagnostic values and M3 had a moderate diagnostic value. Conclusions The RF method can be used as a method of tongue image recognition in CHD with Qi deficiency and blood stasis syndrome. The characteristic values of M1 , M2 , M3 and M5 can be used as specific indications of clinical samples and animal models of CHD with Qi deficiency and blood stasis syndrome. The rat model of CHD Qi deficiency and blood stasis syndrome is in line with the digital characteristics of tongue diagnosis in patients.