The paper proposes the coupling method based on Bayesian formula and fuzzy recognition by improving the Bayesian classification algorithm. The combination weighting method and relative membership degree are introduced to clarify the weight of evaluation index, and the maximum likelihood classification criterion is used to determine the grade of individual primary school students' physical health evaluation index. This method is applied to physical-health evaluation of six students in Shuangfeng primary school in Jiujiang City. The results show that some student VI is worse in physical health evaluation, the other students are medium or above, which is consistent with the actual sampling results. The method achieves 67% comprehensive accuracy and has certain credibility, which realizes the correct judgment on physical condition of primary school students in probability, and take the initiative to warn the primary school students who have a higher probability of physical hidden-danger. In this case, group physical judgment and personalized effective intervention are carried out on students to promote their healthy development and the overall physical level of primary school students.

通过改进贝叶斯分类算法，提出了一种基于贝叶斯公式和模糊识别的耦合方法。引入组合加权法和相对隶属度，明确评价指标的权重，采用最大似然分类准则确定小学生个体健康评价指标的等级。该方法用于九江市双峰小学六名学生的身体健康评价。结果表明，部分学生VI的身体健康评价较差，其他学生中等或以上，与实际抽样结果相吻合。该方法综合准确率达到67％，具有一定的可信度，从而实现对小学生身体状况概率的正确判断，并主动警告那些有较高身体隐患危险的小学生。在这种情况下，将对学生进行集体身体判断和个性化的有效干预，以促进他们的健康发展和小学生的整体身体水平。