<p>Teachers’ technological knowledge is essential for integrating Artificial Intelligence (AI) into language education. Additionally, teachers’ self-efficacy and attitudes toward AI can foster the integration of AI tools in education. This study developed an Artificial Intelligence – Technological Pedagogical Content Knowledge (AI-TPACK) framework to measure the specific knowledge and skills required for pre-service teachers specializing in teaching Chinese as a second language (TCSL) to effectively integrate AI into their teaching practices. A total of 520 pre-service TCSL teachers participated in the study. The development of the AI-TPACK framework involved Explanatory Factor Analysis (EFA) with 284 participants and Confirmatory Factor Analysis (CFA) with 236 participants to evaluate the model fit. The finalized framework consists of 25 items, encompassing seven dimensions: Content Knowledge (CK), Pedagogical Knowledge (PK), AI-Technological Knowledge (AI-TK), Pedagogical Content Knowledge (PCK), AI-Technological Content Knowledge (AI-TCK), AI-Technological Pedagogical Knowledge (AI-TPK), and the synthesized form AI-TPACK. Furthermore, Structural Equation Modeling (SEM) was used to examined how pre-service TCSL teachers’ self-efficacy influences their AI-TPACK, focusing on the mediating role of AI attitudes. The results demonstrated that pre-service TCSL teachers’ self-efficacy significantly influences their AI-TPACK through direct and indirect pathways, with AI attitudes serving as the mediating factor. This study made several recommendations for future professional development programs, emphasizing the need to bolster teachers’ self-efficacy and enhance their positive attitudes toward AI to better prepare them for the evolving language educational landscape.</p>