Objective To apply and evaluate domestic intelligent chatbots for psychological counseling, and systematically assess their strengths and limitations. Methods Intelligent chatbots supporting psychological counseling were systematically retrieved from the Apple Store and WeChat Mini Programs, and all tools meeting the inclusion criteria were downloaded one by one.Three simulated cases were designed to interact with each chatbot in sequence, generating manuscripts with a total word count of 240,000. The Chatbot Usability Scale-15 (BUS-15) was adopted to evaluate the chat function quality of each tool, and the Behavior Change Technique Taxonomy was used to analyze the psychological intervention techniques adopted by each chatbot. Results A total of 34 AI chatbots were included. The inter-rater reliability for BUS-15 between the two evaluators was 0.793.The average scores of all chat tools on each dimension of the BUS-15 ranked from high to low as follows: privacy, perceived accessibility of chatbot functions, response time, perceived functional quality of chatbots, and perceived quality of communicated information. The usability scores of chatbots powered by large language models were significantly higher than those of professional psychological counseling chatbots and chatbots without large language model support (both P<0.05). Each chatbot adopted an average of 5.6 behavior change techniques. Conclusion Current professional psychological counseling chatbots need improvements in communication flexibility and privacy protection, while general-purpose chatbots should strengthen their professional guidance capability. Several tools require optimization in information volume, elimination of potential discriminatory expressions and ambiguous wording. Widespread adoption of large language models to optimize communication styles will be a major trend in the future.