This paper presents an approach to reconstruct 3D objects using Genetic Algorithm (GA) from a 2D image. In the proposed method GA evolve 3D initial models to a target object by means of comparisons between images generated from the models and image of target object. Promising results were achieved reconstructing simple 3D geometric objects. This technique is useful for creating random model by using 3D primitives. For initial experiments we have selected sphere, cube, and cone. The experimental results shows that proposed method generate 3-D model automatically as compared to conventional methods for generating 3-D models. © 2010 IEEE.