Aesthetic literacy originates from rich aesthetic experiences, one of the principal ways people gain such experiences is through arts appreciation. Sixty-seven students participated in 119 art appreciation activities at out-of-school public art venues in China. A digital education platform was used to record students' appreciation process. A recommender system was used to recommend relevant art knowledge to students. The data obtained were used to analyse the learning characteristics of students in out-of-school art appreciation and the effectiveness of the recommender system in forming aesthetic experiences. The analysis showed that students were satisfied with this learning model, with NPS = 37.313, but their self-regulation was low (k = 0.606) and universal (std = 2.755). The efficiency of the recommender system was significantly associated with several aspects of students' motivational learning strategies, including intrinsic goal orientation (0.503), task value (0.638), and critical thinking (0.643). Focus group interviews of students with different characteristics revealed that the autonomous and interactive nature of out-of-school art appreciation was the main attraction for students and the recommender system was able to complement and improve students' art knowledge and broaden their artistic horizons. Overall, in an informal learning environment, the use of an AI recommender system can effectively enable students to form aesthetic experiences.