A new method for merge of range data and texture based on calibrating two cameras in a world coordinate system is proposed. The overlapped zones and holes of data of the cameras are detected by use of modulation as mask. The part with specified size in the overlapped zones is used for data integration, according to the nearest sequence number of samples in unit area. The method for merging regions where there are holes in one side and none in the another side is also introduced. Integrated region is constructed with the overapped zones by enlarging the edge of the hole. The data in all integrated zones are averaged with distance to their center line as weight number. Thus, the data transition of each side is smooth. The experimental results reveal this method is reliable and accurate.