In the field of Human-Computer Interaction (HCI), the application of anthropomorphic design in AI doctor consultation systems is gaining increasing attention. Anthropomorphic design elements, by endowing AI systems with human-like characteristics, can enhance users’ attention of AI doctors, thereby influencing their perceptions and behavioral intentions. In HCI research, anthropomorphic design has been shown to improve users’ trust, acceptance, and emotional experience. This is especially true in telecare, where anthropomorphic design enables AI doctors to appear more empathetic and approachable, effectively bridging the gap between humans and machines. This study aims to explore the effects of AI doctor appearance (anthropomorphic vs. real human) and the form of representation in text-based dialogues (avatars vs. virtual characters) on users’ perceptions and behaviors. The experiment recruited 40 participants through convenience sampling. The results reveal the following: (1) A real human appearance helps enhance users’ perception of the AI doctor’s image. (2) Anthropomorphic appearance increases users’ expectations and intention for continued use. (3) When the form of representation in text-based dialogue is a virtual character, users’ perceptions and behavioral capabilities are significantly affected. (4) We found significant interaction effects on both expectation and intention to continue using: when the AI doctor was represented anthropomorphically, users’ expectations and intentions to continue using the virtual character were higher than with the avatar form; moreover, when the AI doctor was represented as a real person, users showed higher expectations and intentions to continue using the avatar form rather than the virtual character. These findings not only provide important theoretical support and practical guidance for the design of AI doctor consultation systems but also offer valuable insights for Human-Computer Interaction design in other application scenarios.