The purpose of this research is to investigate traffic accidents causes and major factors that would be the main reasons for accidents, and to develop a model for accidents on Main National Egyptian Rural Roads. Accidents models would help decision makers to develop the road network to be more safe and minimize the accidents rate. Models were calibrated using accidents records data collected from 19 rural roads from previous studies, and 7 rural roads from filed surveys in Assiut region, "General Authority for Roads & Bridges". The data was splittied in two different types of road sections in location and engineering properties, namely; "Agriculture roads undivided sections", "Desert roads undivided sections”. Simple, Stepwise, and multiple regression analysis have been used to find the effect of each parameter on the accident rate value. Several functional forms are explored and tested in the calibration process. Before proceeding to the development of models, ANOVA statistical tests are conducted to establish whether there are any significant differences in the data used for models’ calibration as a result of differences among the considered roads. The results indicated that exponential model formula represents the highest correlation for all road types .Average daily traffic (ADT), Shoulder width, pedestrian crossing and percentage of trucks have highest effects of traffic accidents on agriculture roads undivided sections, Shoulder width, Radius of horizontal curves and percentage of trucks have the highest effects of traffic accidents on desert roads undivided sections.