This research aimed to compile a leisure involvement scale for in-line hockey players with high reliability and validity by employing a series of well-organized analyses. The athletes who participated in the sixth national in-line hockey match were taken as the sampling population and a total of 250 valid questionnaires were collected. The collected data was analyzed by adopting CFA (confirmatory factor analysis, a statistical technique under SEM, or structural equation modeling) and the compiled scale was proved to have high levels of convergent validity, discriminate validity, and composite reliability. The leisure involvement scale compiled, as the research result shows, was a first-order, multifactor model with three variables-pleasurably, centrality, and level of symbolic meaning.