Since 3D models have been used in wide range of industries, such as entertainment, architecture, medical, etc, the availability of tools to ease the creation of 3D models is important. This paper proposes and implements an algorithm for converting 2D image into sequence of curves on 3D flat model. This work is a part of a system for applying detail features captured from 2D image into 3D model. With 2D image as the input and 3D flat model containing sequences of curves as the output, the main processes in our algorithm are edge detection, connected-component labeling, finding list of control point's curves, mapping 2D canvas into 3D flat model, and drawing cubic B-spline curve which interpolates endpoints. The experiment shows that the curves on 3D model can approximate the shape-content of 2D image, and the processing time of algorithm is sufficiently fast.