Objective To explore the application effect of artificial intelligence (AI)-powered multimodal teaching evaluation in the teaching practice of infectious diseases, and to evaluate its impact on the learning effect of clinical medical students. Methods A total of 180 students from Class of 2021 majoring in clinical medicine at Youjiang Medical University for Nationalities were randomly divided into an observation group and a control group, with 90 students in each group. The observation group adopted AI-powered multimodal teaching evaluation method, including an AI-assisted virtual case analysis, online interactive discussion platform and real-time feedback mechanism; the control group received traditional teaching methods. The teaching period lasted 16 weeks, and the learning effects of the two groups were comprehensively evaluated by means of theoretical examinations, clinical skill operation assessment and student satisfaction questionnaires. Results The total scores of theoretical examination, practical operation and all individual modules in the observation group were all higher than those in the control group. The average online learning time in the observation group was (5.67±1.34) h/week, the knowledge revisit rate was (9.23±2.05) times/week, and the human-computer interaction frequency was (13.45±3.01) times/class. The learning behavior showed the characteristics of “high frequency and short time”, and the knowledge revisit was concentrated in the 24 hours before class. The revisit rate of knowledge in the observation group was significantly positively correlated with case analysis, comprehensive application and virtual case processing; the frequency of human-computer interaction had the strongest correlation with clinical skill operation and virtual case processing; the learning duration was only weakly correlated with knowledge mastery. The satisfaction of the observation group was higher than that of the control group in terms of technical ease of use, learning participation and depth of knowledge mastery, but 13.33% of the students in the observation group reported that they had difficulty in technical adaptation in the initial stage. Conclusion AI-powered multimodal teaching evaluation has significant advantages in the teaching practice of infectious diseases, can effectively improve the comprehensive learning effect of students, can be used as a powerful supplement and innovation of traditional teaching methods, and is worth further promotion and application in the field of medical education.