Neurolinguistics has been widely recognized for its focus on exploring the formation mechanisms of human language and its potentially significant impact on language expressivity. This article introduces an evaluation mechanism based on the BP neural network. It can help us better evaluate teaching effectiveness and potentially significantly influence teaching outcomes. We aim to enhance teaching efficacy and assist learners in improving their learning outcomes. Using MATLAB to assess the teaching efficiency of university English classes, we found that this approach eliminates the subjective influence of professionals and yields satisfactory conclusions applicable to various circumstances.