Reasonable calibration can improve vehicle drivability of the mode of engine starts. Aiming at the problem that calibration standard is not perfect, which rising a difficult issue to how to calibrate engine management system reasonably. In this paper, drivability influencing factors of vehicle was analyzed from the control for engine management system under the mode of engine start and engine stop. Furthermore, the evaluation model of drivability of engine starts was established which provided a repeatable and quantifiable method for the calibration of engine management system. The drivability influencing indexes was analyzed from the subjective feeling of drivers and dynamic change on engine starts process, and according to the process of engine start, crank time, start time, return to idle time, overshoot, undershoot and surge were considered. In order to verify the correlation of these indexes and got the key indexes, the Pearson correlation analysis method and the principal component analysis method was used to determine the key influencing factors of the drivability. To achieved drivability evaluation of engine start, the evaluation model of drivability was established which based on support vectors machine theory about engine start. Through compare to the result of subjective scoring and modeling scoring, found that the evaluation model can correctly describe the drivability quality in the mode of engine start. The results derived from the evaluation model can also be used calibration targets to achieve repeatable and quantifiable calibration of the engine management system. This method can give the practical significance and theoretical basis for the calibration work of the control strategy of engine start.