A study was undertaken to develop a methodology to estimate annual average daily traffic (AADT) for nonstate roads in urbanized areas in Florida. The current practice related to the estimation of AADT for nonstate roads has been of great concern to the Florida Department of Transportation because of the potential lack of accuracy in the estimated data. For this study, a multiple regression model was developed for estimating AADT on nonstate roads. The model utilized a large sample size (data from 450 count stations in Broward County) and involved the investigation of up to 12 initial variables. Various methods, including geographic information systems (GIS), were explored to convert the current digital data and aggregate them into suitable forms for statistical analysis. Statistical tests were performed and the results showed that the most important contributing predictors are roadway characteristics, such as the number of lanes, functional classification, and area type. Socioeconomic variables, including population, dwelling units, automobile ownership, employment numbers, and school enrollment in the surrounding area, have an insignificant impact on AADT. Further analyses revealed the deficiency of traditional roadway functional classification, the need to improve the method of road classification, and the need for alternative methods to account for the impact of economic activities on AADT.