During pedestrian road crossing, pedestrian accepts vehicle gaps with different road crossing behavior and has significant influence on design parameters for traffic operations. The present study analyzes the pedestrian accepted gaps (time gaps) with different roadway geometry at unprotected (unsignalized) mid-block crosswalk locations under mixed traffic conditions. For this purpose, eight unprotected mid-block crosswalk locations were selected in Mumbai, India. A Multiple Linear Regression (MLR) model was developed by k-fold cross-validation method. The model results show the decrease in pedestrian accepted gap size due to pedestrian behavior such as rolling behavior and speed change and this behavior of pedestrian is more frequently observed at roadway with lower number of lanes as compared to the six-lane divided roadway. Further, increase in vehicle speed, number of lanes, and pedestrian age significantly increases the pedestrian accepted vehicle gap size. The study findings may be useful in pedestrian facility design or developing crosswalk warrants.