Clinical decision support system (CDSS) is one important application for artificial intelligence. Its utilization in clinic evokes controversial argument. Obviously, its development is not as good as expected to reduce the workload of clinicians, to improve the accuracy of clinical decision, to reduce misdiagnosis etc. Here arises the calling on ideology innovation for building CDSS to improve its performance in clinic. Being aware of that problem-solving in clinic is a prerequisite for CDSS to complete its tasks, issues from six aspects that CDSS must address were identified, which might help to find a right way to construct CDSS. It is more than a technic work, it is the cognition over CDSS and requires effort from clinicians, software scientists, and medical scientist together.