The rapid population aging problem and the comprehensive medical insurance system in each country exacerbate the burden on healthcare institutions. Cutting down patients’ waiting time and seeking reasonable distribution of doctor’s idle time during clinic operations are two issues of concerns. This study investigates two appointment scheduling patterns, which are used to compute patients’ suggested arrival time, and three late arrival policies, which are used to determine the service order for patients who miss their assigned numbers, in order to find a best combination of them for each particular clinic department. This study assumes that patients’ decision of arrival time is not a fixed parameter. Instead, it is a variable changing with patients’ number of visits. Patients would adjust their preferred arrival time continuously using incomplete information and previous clinic experiences according to the adaptive expectation theory. We apply a two-stage decision model in which clinic choses one appointment system in the first stage, and then patients decide their arrival times accordingly. The results reveal that the “plateau-dome” scheduling pattern and three-queue late arrival policy reduce patients’ waiting time more effectively than other strategies. Moreover, distribution of doctor’s idle time is more reasonable when applying “plateau-dome” scheduling pattern than that when using the other appointment scheduling pattern.