Shot peening is an important mechanical treatment process for metal surface. Its treatment quality is closely related to the shot motion behavior which is derived by high speed airflow. To accurate control shots motion and improve treatment quality in shot peening, it is great significance to explore an accurate measurement method and uncover the law of shots motion in peening process. With high speed camera, a high-speed image data acquisition system for shots in peening process was established. In addition, the images of instantaneous spatial position of shots in peening were captured. In order to improve contrast, Photoshop was used to covert the collected photos into black-and-white binary images. The open source particle image velocimetry（PIV） software PIVLab was used to analyze these images, and the velocity field of shot flow was carefully calculated. The results show that, under the condition of 0.4 MPa gas pressure and 0.4 mm cast steel shot, the scattering cone angle of shot flow was 10 degrees, and the maximum velocity of shot was about 46.5 m/s. On different spatial sections, the velocity of shots presents Gaussian distribution. In addition, the standard deviation of this Gaussian distribution is increased with enlarge the distance between shots and nozzles. The velocity of shots increases rapidly to a stable value after leaving nozzle, and the velocity variation conforms to the Logistic function. Compared with the empirical formula, the obtained formula of velocity field distribution of shots is more accurate. Furthermore, the spatial distribution characteristics of shots flow in peening process can be exactly gained.