This paper proposes a new forecasting approach for telephone network routing plan byapplying Neural Network concept. For the experimental results, traffic data from Khonkhaen1 (KKN1) and Khonkhaen3 (KKN3) exchanges
are collected and compared with the forecasted values by using back
propagation neural network method and the Conventional Method. This paper shows that the forecasting method by applying neural network which is convenient and suitable fornetwork routing plan. This method can effectively forecast the traffic flown between the exchanges, moreover, dimensioning for each future network routing plan can be done appropriately.