Although there is an issue with erroneous result evaluation, oral teaching assessment plays a crucial role in teaching higher vocational English oral communication skills. When it comes to higher vocational English oral instruction, the conventional learning score approach is both ineffective and unfair in its evaluation of students' progress. So, to assess and analyze spoken instruction, this article suggests a deep learning scoring model. In order to decrease the interference element in the assessment of oral teaching, the indicators are split according to the needs of oral teaching evaluation and the deep learning theory is utilized to assess the teaching. The next step is to use deep learning theory to create an assessment system for oral instruction of higher vocational English and then to combine the evaluation findings. Based on the results of the VIVIDO simulation, the deep learning scoring model for oral instruction in higher-level vocational English meets certain assessment requirements. Traditional learning scoring methods were outperformed by oral teaching assessment in terms of both accuracy and evaluation time.