An improved inertial and visual sensors calibration method is proposed in this paper, which is the method of fractional step to calibrate between inertial and visual sensors. The relationship of the rotation between inertial and vision sensors can be obtain by solving the improvement of hand-eye calibration equations. The inclinometer sensor is introduced in the process of calibration and the output of inclinometer data is utilized to optimize the rotation matrix between inertial and vision sensors; the translation between inertial and vision sensors can be obtain through solving the basic hand-eye calibration equations. The experiment results demonstrate that the method can precisely calibrate the relative position of the space between inertial and visual sensors.