In order to improve the visual effect, we can use the intelligent analysis method to carry out deep active learning, verify the effectiveness of the research results, show the deep learning method, can judge the intonation and timbre in speech recognition, complete the comprehensive analysis, and improve the overall analysis effect. The recognition accuracy rate reaches between 80 and 90%, and the simplification rate of English speech recognition is simplified, with a simplification rate of 23%, so the deep machine learning method can improve the speech recognition effect and promote its application range.