Applications based on Wireless Sensor Networks (WSN) are influenced by many factors such as transmission errors, network topology and power consumption. Consequently, developing such applications introduces several research challenges. In this paper, we propose an intelligent model in order to achieve energy efficient message forwarding over a WSN. Based on the proposed model, a network node evaluates a specific statistical extrapolation error through local linear regression and Lagrange polynomial methods and adaptively decides whether message retransmission is required or not. Simulation results are reported denoting the applicability of the proposed model in minimizing energy consumption on message forwarding.