Automatic programming is an efficient technique that has contributed to an important development in the artificial intelligence field. In this paper, we introduce a new technique called Variable neighborhood Programming (VNP) that was inspired by the principle of Variable Neighborhood Search (VNS) algorithm. VNP starts from a single solution presented by a program, and the search for the good quality global solution continues by exploring different neighborhoods. The goal of our algorithm is to generate a good representative program adequate to a selected problem. VNP takes the advantages of the systematic change of neighborhood structures within a local search of the VNS algorithm to explore more research space. To explain more the algorithm process we apply VNP in a simple sample in symbolic regression problem. Then the effectiveness and the good convergence of this algorithm is proved by testing it on benchmark problems drawn from time series prediction and classification areas, and we compared it with the related techniques.