The intelligent speech training system aims to use real-time speech recognition to analyze whether the user’s speech contains too many filler words and derogatory words. In addition, facial expressions are also an indispensable part of the speech. The system can analyze whether the user’s expressions are appropriate, so as to achieve expression management. When the standard is not met in the speech, the system will also use a warm reminder to let users know their shortcomings in real time. Furthermore, the system also comes with the functions of replaying video and viewing reports, which can be sent to experts for further evaluation. Finally, what is indispensable is how to prepare a high-quality and impressive report file in advance. The system can help users quickly analyze the word count, picture, and font size of the report file, and quickly give users suggestions for modification.