<p>With the rapid development of artificial intelligence, digital technology-based error correction significantly improves the efficiency of error correction by automatically identifying and clustering students’ errors and then analyzing the types of errors. However, in China’s whole class teaching, teachers still rely on experience to randomly select students with representative errors for error correction, making it difficult to ensure that all students’ errors are corrected in a timely and effective manner. In addition, in Chinese primary classrooms, high-achieving students and low-achieving students learn together in the same class, making it difficult to ensure that each student achieves the learning goals. Therefore, this study proposes a peer teaching-based digital error correction approach, focusing on its effects on primary school students’ learning performance, learning engagement, and perceptions of error correction. A total of 63 primary school students were recruited for the study, with 31 in the experimental group using the PT-DEC approach and 32 in the control group using the E-DEC approach. The results showed that students using the PT-DEC approach performed better than the control group in terms of learning performance and learning engagement. The results of this study validate the effectiveness of peer teaching in digital error correction and provide valuable insights and guidance for exploring more efficient error correction in the future.</p>