Road transportation systems have undergone a significant transformation in the recent years. Emphasis is no longer solely on building roads to interconnect various places, cities or states. Other factors such as real-time navigation, safe and clean driving, travel comfort, driver assistance, quick reaction to avoid occurred incidents, and optimal usage of road transport infrastructure are becoming more and more significant. Hence, to ensure efficient usage of the traffic system capacity, intelligence needs to be added to its control systems. Such control systems enable an autonomous response of the traffic infrastructure to changes in the traffic flow or wheatear conditions, such as activating service lanes as full road lanes to increase maximal possible traffic flow, activating dynamic traffic signs to inform drivers about new road conditions, rerouting vehicles to avoid an incident situation, etc. One of the elements that gained a significant influence on a road capacity and travel times are tollbooth plazas. They are a possible bottleneck and can reduce travel comfort. This paper presents a fuzzy logic-based approach to control the number of active tollbooths in order to ensure desired Level of Service (LoS) with minimum of expenses. Input variables for proposed controller are current queue length and its change in two consecutive time steps. Controller output is the value that determines the type of change (increase, decrease or no change) of the active tollbooths number. A quality criterion for tollbooth plaza capacity control is proposed regarding to desired LoS defined with minimal and maximal waiting time constraints. Proposed criterion is used to evaluate the implemented fuzzy logic controller. In order to enable testing of the fuzzy logic controller a stochastic queuing theory model, for the tollbooth plaza capacity, is implemented as a Matlab/Simulink simulation model. The model enables recording of all relevant control and tollbooth plaza model variables such as number of vehicles in a queue (cumulative and per active tollbooth), waiting time, tollbooth plaza traffic load, and number of active tollbooths. Usability of implemented model and controller is also examined for system behaviour prediction by using estimated values of traffic flow density. In order to provide relevant simulation results real traffic flow data are used including real toll payment time constants.