With the continuous development of the computer field, mixed reality technology is increasingly used in various fields of life. This paper aims to create a mixed reality interactive experimental teaching system for economic colleges and develop the overall framework of the system using the target detection algorithm. It first designs the interaction system module based on monocular visual spatial localization. Then it builds a teaching evaluation system based on face detection, recognition, and gesture estimation for the experimental teaching of economic colleges and universities for the mixed reality interaction system. This paper empirically analyzes the proposed mixed reality interaction system for experimental teaching in economic colleges. It can achieve a processing speed of 19 frames per second and a high recognition accuracy of 88.23% in the teaching test. The average scores of each index after the testers’ use are above 4, and the average value of the system’s overall score is 4.66. This study has excellent performance and provides a practical roadmap and reference value for developing mixed reality technology in the direction of experimental teaching.