With the improvement of quality of life, people pay more and more attention to the comfort performance of clothing, of which thermal and wet comfort is an important part of evaluating the comfort of clothing, referring to the performance of keeping the human body in a reasonable thermal and wet state. When the human body sweats a lot or is in a highly humid environment, the clothing fabric will be soaked to make people feel wet, which seriously affects the comfort performance of clothing wear, and with the rapid development of sensing technology, the comfort of human clothing can be comprehensively evaluated by a variety of sensing data (clothing pressure, temperature, humidity, and heart rate). Therefore, how to analyze and process these data and establish an objective and accurate evaluation criterion for clothing comfort is a difficult problem and has attracted the attention of many researchers. In this paper, an improved kernel function fuzzy kernel c-means clustering algorithm is used to analyze the pressure at specific points in human activities. Unsupervised clustering analysis was performed for five clustering metrics (mean, pressure range, temperature range, humidity range, and heart rate variability). The clustered samples were learned and discriminated by a support vector machine to determine the comfort level of the clothing. The method can be applied to multi-indicator and multiclassification problems, providing smart clothing researchers with an intelligent, objective, and accurate method for evaluating clothing comfort. Experiments show that the method designed in this paper has good performance experience in terms of mean value, pressure range, temperature range, humidity range, and heart rate variability.