Breast cancer has a high metastasis rate. Quantitative analysis of breast cancer growth and metastasis by real-time imaging using small animal imaging technology can help understand the disease mechanism and help drug research. Two dimensional imaging can not achieve high precision positioning and quantification of optical signals. With the progress of computer technology, it can realize the three-dimensional reconstruction of the collected images, accurately quantify the optical signals and obtain the three-dimensional information of spatial distribution. IVIS spectrum small animal in vivo optical three-dimensional imaging system has high sensitive bioluminescence, fluorescence, cherenkov radiation two-dimensional imaging and three-dimensional scanning reconstruction functions. It is the top system of small animal in vivo optical imaging. In this study, human breast cancer cells (MDA-MB-231) were infected with lentivirus. After stable expression of luciferase, the animal models of severely combined immunodeficiency (SCID) were established. The bioluminescence two-dimensional imaging of the mice was carried out through the IVIS Spectrum living animal optical three-dimensional imaging system. The growth and metastasis of the tumor were observed noninvasively. The innovation of this paper is to make quantitative three-dimensional imaging of breast cancer mice by bioluminescence tomography. The algorithm is directly used for 3D reconstruction with the algorithm of the system. The network model of the deep encoding decoder combined with the optimized convolution of WOA is reconstructed. The reconstruction effect is verified by CT diagram, and the depth information of the tumor is obtained, and the accurate quantitative analysis of breast cancer is achieved.