Many contemporary fueling centers are much larger and offer more services than traditional gas stations. For these fueling stations, the information contained in the Institute of Transportation Engineers (ITE) Trip Generation Handbook for gasoline stations may not be applicable. However, quantifying trip ends from fueling centers is very difficult due to the many site characteristics that influence whether or not a fueling center is chosen. This paper highlights a study that was conducted to investigate the importance of site and variable selection in developing an improved trip generation model for fueling centers. Thirty data collection sites were studied. Primary site selection factors included the presence of at least 8 fueling positions, the presence of a fast-food restaurant and geographic location across the state. Trip generation data were collected over a three-month period. Two approaches were used for analyzing the data: the linear regression approach typically used by ITE and a multi-variable linear regression analysis. The results showed that the multi-variable equations were able to represent the fueling centers more precisely that the single-variable approach. The recommended trip generation model uses adjacent traffic volume and type of services rather than size attributes in trip generation.