With students in civil engineering often having limited programming experience and facing difficulties in transforming complex engineering problems into computational programs, this study explores an AI-assisted programming teaching practice based on the strip method calculation task in the course Bridge Structural Seismic Design. A teaching process integrating engineering problem analysis, computational model construction, and AI-assisted programming was developed. The effectiveness of the teaching approach was evaluated from students’ learning processes, consistency of computational results, and feedback. After two rounds of teaching practice, the results showed that AI-assisted programming reduced the difficulty of program implementation and helped students solve complex engineering problems through computational methods. However, students still exhibited deficiencies in model construction, appropriate abstraction, and program debugging. Therefore, the application of generative artificial intelligence in civil engineering education should emphasize human–AI collaboration, encouraging students to continuously analyze, modify, and validate programs, thereby promoting the development of computational thinking and human–AI collaborative abilities.