The primary objective of this study was to develop an automatic system for summarizing customer reviews in Thai. it is the system that helps summarizing customer review text in e-commerce websites. This system considers only features of the products and whether the opinions towards the products in the reviews are positive or negative. A system for summarizing customer reviews in Thai has not been widely recognized. Thus, in this research, we applied techniques such as word segmentation, creating and orientation identification for opinion words in seed list, feature extraction and orientation identification for feature. Which are composed of the customer review summarizing system and studied related configurations such that the developed system has a satisfying performance on Thai. Our task is performed in three steps: (1) Collect and preprocessing reviews; (2) Extract product features from the reviews and create list of opinions together with their orientation. (3) Identifying orientation of each features from opinion words in the reviews and summarize the results. The developed system was tested on reviews that were collected from a cosmetic selling website including 1,680 customer reviews which can be divided into 4 categories. The results are compared against the results summarized by marketing experts. The results show that the recall in identifying product features is 0.77 and the corresponding precisions before and after word filtering are 0.33 and 0.53 respectively. In addition, the accuracy in identifying the orientations of customer opinions towards product features is 0.58.