In order to reduce the system errors of dead reckoning and improve the localization accuracy,a new model for systematic error of mobile robot was defined and a UMBmark-based method for calibrating and compensating systematic error was presented. Three dominant reasons causing systematic errors were considered: imprecise average wheel diameter,uncertainty about the effective wheelbase and unequal wheel’s diameter. The new model for systematic errors is considering the coupling effect of the three factors during the localization of mobile robot. Three coefficients to calibrate average wheel diameter,effective wheelbase,left and right wheels’ diameter were obtained. Then these three coefficients were used to make improvements on robot kinematic equations.The experiments on the dual-wheel drive mobile robot DaNI show that the presented method has achieveda significant improvement in the location accuracy compared with the UMBmark calibration. In order to reduce the system errors of dead reckoning and improve the localization accuracy, a new model for systematic error of mobile robot was defined and a UMBmark-based method for calibrating and compensating systematic error was presented. : imprecise average wheel diameter, uncertainty about the effective wheelbase and unequal wheel’s diameter. The new model for systematic errors is considering the coupling effect of the three factors during the localization of mobile robot. Three coefficients to calibrate the average wheel diameter, effective wheelbase, left and these three coefficients were used to make improvements on robot kinematic equations. These experiments on the dual-wheel drive mobile robot DaNI show that the presented method has achieved significant improvement in the location accuracy compared with the UMBmark calibration.