Computer system ability (CSA) can enable students to design and analyze a computer problem from a system perspective, understand the integrity, hierarchy and relevance of different modules within a computer system, and master the correlation between computer hardware and software. Since 2015, the CSA training proposal has been widely promoted in many Chinese universities, and has effectively improved the quality of college computer education. However, with the development of computer technologies such as artificial intelligence (AI) and big data, the traditional CSA training proposal cannot fully meet the needs of future computer education. To address this challenge, this paper improves and extends the currently widely used CSA training program. On the one hand, the traditional CSA proposal is improved by integrating the knowledge of more computer science (CS) courses and requiring students to build a more complex computer system. On the other hand, the traditional CSA proposal is extended by proposing a new additional CSA training program dedicated to AI professionals. The above improved and extended CSA training proposals have been implemented in the School of Computer Science and Artificial Intelligence of Wuhan University of Technology for more than five years, and the teaching quality evaluation results show that these proposals have effectively improved the quality of talent training in CS specialty.