In this paper we have proposed a minimum noise shortest path determination scheme considering the amount of delay and energy consumed with respect to each path. An artificial neural network has been employed for classifying the minimum noise shortest path from the source to destination. A simulation work has been carried out with respect to different Signal-to-Noise (SNR) values in a thirty-node network with one Internet node and 100 bits of message length. Also, a comparison has been made between plain Dynamic Source Routing (DSR) and integrating the minimum noise shortest path algorithm with DSR. The simulation results show that with the increase of SNR, noise constraint in the path reduces, and data throughput increases.