A Two-Part Negative Binominal distribution (Two-Part NegBin) model is used in this study to analyze the decision to utilize outpatient care under National Health Insurance (NHI). The decision making process is separated into two parts: the contact analysis and the frequency analysis. The empirical results show that most factors have inconsistent impacts on the contact and frequency decisions except for the health conditions. For example, age affects an individual's contact decision; however, copayments and distance affect his (her) frequency decision. The different empirical outcomes between the contact and frequency decisions support the formulation of two decision processes in the outpatient care utilization instead of adopting one-step decision process. Meanwhile, since outpatient care includes three different types of services, which are Western medicine, dentistry, and Chinese medicine, we divide total outpatient visits into three categories in the analysis of the demand for outpatient care. In order to avoid biased estimations and to provide more precise policy implications, it is more appropriate to separate the three categories of services in analyzing patients' utilization behaviors.