Coronary heart disease is the one of the most common diseases and a major cause of death internationally. The early detection and prediction of such disease is thus very important for human life. Currently, the Adaptive Neural Fuzzy Inference System (ANFIS) is increasingly becoming popular in the field of prediction and diagnosis of medical disease, because ANFIS can arrive at the definite conclusion by dealing with ambiguous, imprecise and vague information in activities or processes. This paper reviews the application of ANFIS in the field of heart disease prediction, as well as some innovative combinations of ANFIS and other techniques for clinical decision support on heart disease diagnosis. Finally, we identify ideas for future work aiming to improve ANFIS model.