Leakage control and monitoring problems of water supply pipe networks are common. The traditional method of partitioning networks based on pressure sensitivity and deploying pressure sensors only considers the pressure changes of network nodes, without combining the elevation information of the network itself, and the effect of excessive local pressure during pressure regulation by pressure reducing valves is not obvious. To solve this problem, the pressure sensitivity of network nodes is coupled with elevation difference, and the distance in the objective function of the fuzzy C-means (FCM) algorithm is defined as the sum of feature distance and spatial distance. A new clustering objective function containing elevation information is established to realize node partitioning clustering. The pressure reducing valve is arranged at the entrance of the partition, and the genetic algorithm (GA) is used to solve the pressure behind the valve to achieve real-time fine control of the pressure of each node in the partition. The pressure sensors are arranged in each partition jointly with the intelligent algorithm and the empirical method, and the rationality of the sensor arrangement is verified by the leakage model. The results show that the scheme divided into four areas reduces the leakage rate of the Balerma irrigate network (BIN) to $6.55 \text{\%}$, which is $22.79 \text{\%}$ lower than that of the initial network. The sensors layout combined with the intelligent algorithm and empirical method has a remarkable effect on the monitoring of network leakage, and finally proves that the optimization strategy of network leakage control based on the FCM algorithm introducing elevation information is effective and superior.