Artificial intelligence evaluation index systems not only serve as a weather vane for the innovative development of global AI but also implicitly reflect the national will and strategic interests of their originating countries. This paper begins by reviewing 22 representative AI evaluation indices from around the world. Based on the evaluation subjects and objectives, it constructs a classification framework for AI evaluation index systems, comprising five categories: comprehensive strength, development foundation, R&D capabilities, ethics and governance, and market application. Secondly, the paper provides an analysis of the differences among these various evaluation index systems in terms of their goal-setting, areas of emphasis in their indicators, and methodologies, revealing the underlying national strategies, policy orientations, and values. Finally, it systematically analyzes four major trends in current global AI evaluation index systems from the perspectives of evaluation paradigms, the right to discourse in evaluation, evaluation objectives, and the timeliness of evaluation.