The traditional method of correcting oral English pronunciation is based on manual experience. However, this method has the problem of low accuracy in correcting oral English pronunciation, and the actual application effect is not ideal. The purpose of this article was to provide an efficient and accurate pronunciation correction method through graph neural networks, enabling learners to better master oral English pronunciation and improve communication skills. Graph neural networks have utilized the characteristics of graph structures to accurately analyze and correct learners' pronunciation problems, helping them better improve their oral expression abilities. The graph neural network-based pronunciation correction model has provided a new pronunciation correction method for college English oral learners. The experimental results of this article indicated that the number of students with pronunciation accuracy, fluency, and expression ability scores above 82 after manual correction was 14, 13, and 10 fewer than those with pronunciation accuracy, fluency, and expression ability scores above 82 after graph neural network correction, respectively.
