This paper uses signal detection and estimation theory to establish a model of students' attention mechanisms in the classroom. Firstly, this paper introduces several commonly used methods for locating signal sources indoors: the nearest neighbor method, weighted K-nearest neighbor method, and approximate method. Then, the algorithm principle is introduced in detail. By comparing the classroom-measured data with the simulated values, the teacher found that the fusion model has a good effect on detecting students' attention. The research in this paper can be used to improve students' concentration and indoor signal detection.