Sign language recognition is emerging as a vital component of our smart life. In addition, commercial RFID shall become a popular technology for sign language recognition. As we all know, there are 70 million deaf people using sign language as their first language and sign language can facilitate communication with deaf people. However, most of the researches are isolated word recognition. There is few researches about sentence-level sign language recognition. More importantly, they are limited and it is difficult to achieve the desired results of realworld applications. So this paper introduces the first sentence-level sign language recognition system based on RFID. It mainly collects the phase sequence of signals received by commercial RFID device. We obtain relatively pure phase characteristics and present a method to carry out sign language segmentation. Effective feature extraction and classifier selection are crucial to sign language recognition. By evaluating our system in real-word environment, we fill in the gaps between corresponding low-cost sentence-level sign language recognition. We implement and evaluate through extensive experiments and the average accuracy of the method are 96% and 98.11% in different multipath scenarios. The results show that our method has high recognition accuracy and robustness.