The aim of the present research is to model the critical factors of mobile phone use on driver behavior through the exploitation of data from smartphone sensors. Data collected from 100 drivers who participated at a naturalistic driving experiment for six months are analyzed, combined with respective questionnaire answers. Using GLM Poisson regression models, the correlation between driving characteristics recorded by smartphone sensors and the percentage of mobile use (converted to integers) while driving was investigated. Four statistical regression models forecasting the percentage of mobile phone use while driving were developed: one overall model and one per road type (urban, rural, highway). Results indicate that parameters affecting the use of mobile phone while driving are (i) the percentage of driving duration with speed above the speed limit (ii) driving distance, (iii) average deceleration, and (iv) average speed. Across the four models, average deceleration had the most consistent impact, appearing to have a statistically significant negative correlation with mobile use integer values.