| The consumer market are filled with competition now, so consumer need more and more products which have multi-purpose, high grade and the multi-attachments value. The designer not only pay attention to the market development trends, but also pay attention to sensory comfort which the apparel brings to the consumer. In order to enhance the success ratio of the product, it is especially important to establish connection between the consumer's psychological comfortable requirements and the product technical characteristic. By researching the consumer psychological sensory perception, we can confirm consumer's requirements and obtain the products physical performance. Once we obtain the products physical performance, the designer will largely enhance the product success ratio. Once we obtain the products physical performance, we will realize the product technical characteristic, and enhance the technical specification and the manufacture technique to produce the product which is in prospect.First, in this paper eight kinds of textile were selected and made into classical short sleeve shirts. Wearing trial has been applied in chamber which has constant temperature and constant humidity. The multi-attitude scale was applied and tested by gentleman to obtain lots of psychological sensory comfort data. Then, basing on the test data, the subjective data have been proved further by using Kendall' method to ensure the availability of data that was used to establish predict model.Next, three independent factors has been abstracted using the method of classify and factor analysis. These three factors are thermal-moisture factor, tactile factor and pressure factor. Basing on that, a linear model is found between sensory factors and overall sensory comfort using the methods of regression analysis . Contrasting between test data and prediction data, it is found that the linear model has a good prediction ability.Then the new model was established using the methods of artificial neural network, simultaneously many kinds of artificial neural network were compared by. predicting overall sensory comfort to obtain the most superior BPANN. Compared between linear model and BPANN model, it is found that the statistical linear model has clear structure, defined level, bright variable reciprocity, but has no ability to simulate the humanity independently and judge self-sensation independently. Comparatively the BPANN predicting the result overall sensory comfort nearly to the experiment data, but is unable to know the variable relations and the relative importance. Once more, the relationship of fabric performance and sensory comfort has been studied quantificationally using the method of canonical correlation analysis. Using stepwise regression, it is found that objective fabric performance can predict psychological sensory perception .At last, two new kinds of textile were made into short sleeve shirts and the fabric performance and the sensory comfort was tested. Then the sensory comfort prediction model has been proved though new test data. It is found that both linear model and BPANN model have good prediction effect.In a word, a series of flow predicting subjective sensation comfort have been established and three independent factors which impact subjective sensation comfort has been abstracted in this paper. A creative study has been carried in the comparative new field of clothing sensory comfort. The prediction model provide theory support to clothing products design. Using this theory, many kinds of feeling mechanism which consumer have were comprehended, at the same time, the prediction model provide theory support to establish model with statistical or mathematics method and possess some practical guidance significance. |