为探讨大学生锻炼坚持行为的影响因素，本研究基于锻炼认知决策模型，纳入身体自尊这一变量对模型进行扩展，检验增加身体自尊这一变量后，该模型对体育锻炼坚持行为预测和解释力的贡献，以及身体自尊对锻炼承诺和体育锻炼坚 持的影响。 本研究基于锻炼坚持认知决策模型，引入了身体自尊来拓展原始模型，建构假设模型，对原模型和假设模型分别进行检验，对检验结果进行分析和对比。该研究使用横断面调查设计来分发，回收调查问卷，并进行整理和录入。本研究首先对各量表进行信效度分析；运用 SPSS 软件对回收整理的数据进行描述性统计分析和相关性分析；运用 Amos 软件分别对原模型和假设模型进行路径分析，考量模型的拟合优度。通过对数据分析的拟合结果与假设结果进行对比陈述，对假设模型和原模型的结果进行对比，结果显示，本研究中各量表具有良好的信效度；大部分变量之间的相关性显著；验证性分析后，所有因子载荷量都大于.50，处于可接受范围内，C.R.值均大于 0.6，构面的内部一致性较好，AVE 值均大于 0.5，表明本文所选取的研究构面可以较好地解释变量的方差；在原模型中，模型对锻炼坚持行为和锻炼承诺的方差解释率分别为 5%和 50%，拓展后的假设模型对锻炼坚持行为和锻炼承诺的方差解释力分别为 8%和 68%,分别提高了 3%和 18%。 锻炼承诺是锻炼坚持行为的重要预测因素，修订后的模型中，锻炼承诺受到锻炼动机、锻炼效果、锻炼自我效能感和身体自尊的影响，锻炼动机、锻炼自我效能、锻炼效果感知、运动技能、身体状况对锻炼承诺有积极的作用；修订后的模型中，运动技能对锻炼坚持行为有积极的作用；修订后的模型与原模型相比, 拓展后的假设模型提高了对锻炼承诺和锻炼坚持行为的解释率,其中拓展后模型对锻炼承诺的解释率更好；引入身体自尊后，拓展后的假设模型的拟合优度良好。因此，除了通过提高大学生锻炼动机、锻炼自我效能和锻炼效果感知外，还可以通过提升大学生专项运动能力、学习运动技能的能力和对运动竞赛情景的自信心；提升大学生体健康状况、体力、保持锻炼的能力以及对锻炼情景的自信心等来提升大学生的锻炼承诺，从而对促进大学生继续坚持体育锻炼起到一定的积极作用。

In order to investigate the influencing factors of college students' exercise adherence, based on decision-making model to exercise adherence, this study incorporated the variable of physical self-esteem to expand the model. This study examined the impact of physical self-esteem on exercise commitment and physical exercise adherence after increasing the variable of physical self-esteem. Based on the decision-making model to exercise adherence, this study introduced the physical self-esteem to expand the original model and constructed the hypothesis model. And then the study tested the original model and the hypothesis model separately, and analyzed and compared the test results.. The cross-sectional study design was used to distribute, collect questionnaires, and then sorted out and input the data of the questionnaires. Initially, the study carried out the reliability and validity analysis of each scale and performed descriptive statistical analysis and correlation analysis on the collected data buy using SPSS software. Next, The path analysis of the original model and the hypothetical model is carried out by using Amos software, as well as the fit of the model. By comparing the fitting results of the data analysis with the hypothesis results, and by comparing the results of the hypothesis model with the original model, the results showed that the scales in this study had good reliability and validity, and the correlation between the majority of variables was significant. After confirmatory factor analysis, all factor loadings were greater than .50, within acceptable limits, CR values were greater than 0.6, the internal consistency of the facets was better, and the AVE values were greater than 0.5, indicating that the research facet selected in this paper can explain the variance of variables well. In the original model, the variance interpretation rate of the model for exercise adherence and exercise commitment was 5% and 50%, respectively, and the hypothesis model explains the variance of exercise adherence and exercise commitment was 8% and 68%, respectively. Compared with the results of the pre- and post-analysis, the hypothesis model improved the interpretation rate of exercise commitment and exercise adherence, and the hypothesis model had a better interpretation rate of exercise commitment. After introducing the variable of physical self-esteem, the hypothetical model was fitted well. Exercise commitment is an important predictor of exercise adherence. In the revised model, exercise commitment was influenced by exercise motivation, exercise effect perception, exercise self-efficacy and physical self-esteem, which exercise motivation, exercise self-efficacy, exercise effect perception, sport competence, physical condition had positive effects on exercise commitment. In the revised model, sport competence had a positive effect on exercise adherence. Compared with the original model, the hypothesis model improved the interpretation rate of exercise commitment and exercise adherence, and the hypothesis model had a better interpretation rate of exercise commitment. After introducing the variable of physical self-esteem, the fit of the hypothesis model was well. Therefore, in addition to improving college students' motivation, exercise self-efficacy and exercise effect perception, also can by improving college students' special sports ability, the ability to learn sport competence and self-confidence in sports competition scenarios, and by improving college students' physical health, physical strength, and maintenance, so the exercise commitment of college students can be improved. Thus the adherence of physical exercise can get improved.