The deep integration of technology and college sports professional training can effectively stimulate the enthusiasm of students for sports training and, to a certain extent, reduce the danger of professional training. This paper first introduces the virtual reality technology under information technology to build a virtual sports training framework, and through virtual sports training, can effectively realize the immersive training of participants. Secondly, we use LiDAR to obtain point cloud data for the virtual scene of college sports professional training and reconstruct the virtual scene using the Poisson surface 3D reconstruction algorithm. Based on the OpenPose human posture estimation algorithm, the training movements of students in the virtual scene are identified, which helps coaches to better formulate standardized training movements. Finally, to verify the effectiveness of the method given in this paper in college physical education training, a series of quantitative analyses was carried out. The results show that the training loss of the Poisson surface 3D reconstruction algorithm is about 0.087 after 180 rounds of iteration, the verification accuracy is about 87.62%, and the training loss of the OpenPose algorithm is about 0.04% when the number of iterations reaches about 130 times. After using VR technology to carry out professional training in sports, the score of students in the experimental group in terms of active learning interest is 4.487±0.324 points, which is about 1 point higher than that of the control group. Virtual reality technology can effectively promote the innovation and reform of sports professional training in colleges and universities, enhance the student’s interest in learning sports training, and provide a guarantee for improving students’ physical quality.