In this paper, a deep learning ID3 algorithm based on a decision tree generation algorithm is proposed and applied to the comprehensive assessment of English online teaching quality. Secondly, the structure and design method of the data warehouse of the comprehensive assessment software of online teaching quality is also constructed, including the design of the operating system interface and the data warehouse itself. Finally, the application results and analysis of the English web-based comprehensive teaching quality assessment software were conducted through assessment attributes and network training. The results show that the correlation coefficient between the two is 0.9348 from the evaluation results of the sample and 0.9116 from the analysis of the network generalization performance test. The correlation coefficient between the two is close to completely accurate from the evaluation and test results. The software developed in this paper provides a certain reference value for achieving the improvement and optimization of teaching quality.