Abstract The construction of smart classroom is an important symbol of the reform of education informatization, and its application can strengthen and enhance the construction of professional disciplines. The construction of a smart teaching model is done in collaboration with art parametric design and aesthetic assessment of art design works in this paper. The visual salient network at the pixel level of the art design is calculated by segmenting salient objects using the original BASNet model. The diffusion equation is used to calculate the visual position probability map, the probability of the corresponding position is calculated based on the layout of the visual elements, and the optimal layout of the visual elements is generated by the candidate region generation algorithm. By creating an aesthetic evaluation module, the image’s global aesthetic information is modeled, and a mapping relationship is established between the learning image and the aesthetic rating. Finally, an empirical analysis of the smart teaching mode is conducted to investigate the impact of this teaching method on students’ learning efficiency and creative ability. The results show that the comprehensive rating of 100 artworks created under the wisdom teaching mode is 3.09, which is greater than 3. The model is found to be significantly helpful for college art majors, as indicated. College art design teachers can use the research in this paper to develop more effective individualized teaching programs through effective strategic guidance.