This paper brought forward an intelligent computer aided instruction (ICAI) modeling method oriented to parallel robot instruction. According to the learning-origin structure of mechanical specialty students, the study process and cognition characteristics were analyzed, and the comparative research on their quality performance in parallel robot direction was carried out. The ICAI model was established oriented to parallel robot practical instruction combining machine learning and artificial intelligence method. Aiming at the basic instruction requirements of mechanical engineering specialty, the innovation mechanism of this area was researched, and the instruction strategies optimized and realization method were provided. Experiments proved that this method can simulate the implementation course of parallel robot instruction, and improve the instruction effects validly.