Since the traditional Structural Equation Modeling is unable to correctly grasp the preferences and behaviors of consumers at multiple levels, this study combined the virtual reality experience model to analyze the motivations and cognitive behaviors of the public towards sports and fitness. The factor analysis was adopted to measure the latent variables of the virtual reality sports experience function, and then Multilevel Structural Equation Modeling was conducted based on the latent variables to find out the optimal model combination and the best path relationship between people's experience situations and latent variables, to truly grasp the motivations and goals of people's sports experience. The empirical results showed that the main factors of the latent variables of people's satisfaction and loyalty of the group level were different from those of the individual level, which indicated that there was a difference between the group and individual sports preferences in virtual reality sports experience, and it indirectly affected the consumption intention of the group and the individual for virtual reality sports experience.