The present study was set out to assess the contents of polycyclic aromatic hydrocarbons (PAHs) in three types of tollbooths at a highway toll station via direct and indirect approaches. Direct sampling results show that no significant difference could be found in the PAH homologue distributions for samples collected from the car lane/ticket-payment and car lane/cash-payment tollbooths, but both were significantly different from that for the bus/ truck lane tollbooth. The above results could be due to the former two types of tollbooths that were designed for the same type of traffic (i.e., cars and vans), but the latter was designed for a different type of traffic (i.e., buses and trucks). For any given type of tollbooth, the total-PAH content (C(Total-PAHs)) found during the day shift (= 9,370-15,500 ng/m3) were not significantly different from that found during the night shift (= 9,550-14,900 ng/m3), but both were significantly higher than that found during the late-night shift (= 5,560-11,100 ng/m3). During any given work shift we found C(Total-PAH5) for the three types of tollbooths as the following: bus/truck lane (= 11,100-15,500 ng/m3) > car lane/ticket-payment (= 7,260-13,500 ng/m3) > car lane/ cash-payment (= 5,560-9,550 ng/m3). After conducting multivariate regression analyses, we found that none of the three environmental factors (i.e., wind speed,temperature, and relative humidity), except for the vehicle flow rate (Q(Vehicle)) had a significant effect on C(Total-PAHs) for any given type of tollbooth. Considering directly measuring PAH contents was labor-consuming and costly, and the above results suggest the possibility of using Q(Vehicle) to predict C(Total-PAHs) for any given type of tollbooth. After conducting simple linear regression analyses, we found that (1) all resultant regression coefficients were found with positive values indicating that an increase in the Q(Vehicle) would lead to an increase in the C(Total-PHHs). (2) from the magnitude of the resultant regression coefficients indicating that an increase in C(Total-PAHs) caused by per unit Q(Vehicle) for the three types of tollbooths were the following: bus/truck lane > car lane/cash-payment > car lane/ticket-payment, and (3) the resultant R2 values fell to the 0.54-0.75 range indicating that the variations in C(Total-PAHs) could be explained well by Q(Vehicle) for the three types of tollbooths. It is concluded that measuring Q(Vehicle) can be regarded as an effective indirect method for estimating PAH contents in various types of tollbooths.