The skilled machinist can continuously adjust operations in terms of machining efficiency, machinability and cost. This adjustment is done by means of cutting conditions. In the previous paper, the system was proposed to simulate the machinist's thinking process which set the appropriate cutting conditions based on evaluation of cutting state. The process was constructed by fuzzy integral and hill-climbing. But it had become a serious problem that the try counts on searching of the appropriate cutting conditions using the previous system were nearly four times as many as the skilled machinist's. In this parer, it proposed that the skilled machinist's know-how for improvement of searching efficiency is systematized by neural network. This system which is constructed by introducing the neural network into the previous system is superior to the previous system. By using this system, the cutting conditions which can get the high evaluation is obtained, and the try counts was half as many as the counts using the previous system. The accuracy of the system's output and the improvement of searching efficiency were examined by the experiments under various cutting conditions in turning, and the validity of this system was confirmed.