Objective: Compare the depression phenotypes of tumor-bearing mice with breast cancer constructed of two different methods, and the mouse model of breast cancer depression that was more in line with the clinical manifestations and suitable for basic research was preferred.Methods: Constructed a tumor model (4T1 group) with 4T1 breast cancer cells alone and a tumor-depression composite model (4T1+CUMS group) combined with chronic unpredictable mild and unpredictable mild stimuli (CUMS). The experimental period was 42 days, and the weight, tumor size, and survival time of the mice were monitored throughout the whole process, and two depressive behavioral tests (including sugar water preference test, open field test, tail suspension test, and elevated cross maze test) were performed on the 15th and 29th days, respectively. Hematoxylin-eosin (HE) staining was used to observe the pathological changes of hippocampal neurons in brain tissue sections. Results: (1) Body weight: The weight of the 4T1 group and the 4T1+CUMS group began to decrease from 29 days, and the weight of the 4T1+CUMS group was significantly lower than that of the 4T1 group and the control group at the end of the experiment (P<0.001). (2) Tumor size: There was no significant difference in the growth rate of tumor size between the two model groups throughout the experiment (P>0.05). (3) Survival time: the survival rates of the 4T1 group and the 4T1+CUMS group were 100% and 60%, and the first death time of the mice in the 4T1+CUMS group was the 36th day. (4) Depressive behavior test: There was no significant difference between the three groups in the first behavior test (P>0.05), and the two groups showed obvious depressive phenotypes in the second behavioral test. The sucrose preference index and the activity distance in the center area were significantly decreased in the two model groups (P<0.001), and the immobile time was significantly increased (P<0.001). (5) Pathological sections of brain tissue: the number of neuronal cells in the hippocampus of the 4T1 group and the 4T1+CUMS group was reduced, the morphology was irregular, the arrangement between the cells was disordered and the gap was unclear, and some nucleoli were blurred. Conclusion: Although both tumor alone and tumor compound stress stimulation methods can be used to prepare breast cancer depression models, the simple tumor model modeling method is simple, the mortality rate after successful modeling. The long time window is convenient for subsequent drug administration and detection, and the causes of depression phenotype are more in line with the clinical causes and manifestations, which can provide a model reference for future animal experiments on breast cancer tumor-related depression.