Abstract: In order to improve the precision of 3D bone model reconstruction, the CT scanning mode and the bone model reconstruction method using CT images were studied, an inter-slice image interpolation algorithm based on gray feature searching matching point was proposed. The algorithm selects the reference point according to the gray value of the neighboring point of a layer adjacent to the interpolation point; searching for a matching point similar to the reference point gray and the smaller distance in the area of another adjacent layer image based on the reference point; The reference point, the matching point gray value and the distance relationship are used to calculate the interpolation point gray value. Nine hand bone models of human body were reconstructed, and the reconstructed models before and after interpolation were compared with the anatomical models respectively. Results from this study indicate that the mean deviation of average deviation was reduced from 0.30 mm to 0.11 mm, and the mean negative deviation was decreased from 0.27 mm to 0.09 mm.