Tumor image detection is an important technology in the field of medical imaging, which aims to accurately identify and locate the presence of tumors through the analysis and processing of images and provide an important basis for clinical diagnosis and treatment. In today’s society, there is a problem of low accuracy of tumor image detection, which greatly affects the efficiency and accuracy of tumor image detection. The BP algorithm in neural network is an effective tumor image detection technique. In this paper, an image detection system is created by using the BP algorithm, which greatly improves the accuracy of tumor image detection and improves the efficiency of tumor image detection. Finally, the experimental results show that the BP algorithm is easy to operate and the accuracy rate reaches 98.4%.