The work aims to solve problems such as large quantity, many varieties, and difficulty of freight container in current clothing enterprises. The RFID technology was combined with clothing information management. An automatic collection system of clothing information was designed. Through the optimization design of RFID reader antenna number and location deployment of the system, the multi-way tree anti-collision algorithm of RFID tags was improved, so as to improve the reliability of clothing information detection system identification. After calculation analysis, a "M" type four-antenna spatial distribution full coverage detection model was built for the reader. With the cooperation of the clothing detection system identification software, the batch clothing electronic label reliability detection experiment was completed. The maximum clothing packing box detection accuracy reached 100%, and the system detection speed reached 60 m/min. The theoretical identification reached 2 600 pieces per minute. It effectively improves the reliability and collection efficiency of the garment information detection system.