Abstarct: The reduction of potable water resources, combined with the rapid increase in demand due to urban development and increasing costs of water treatment and transmission, has become a serious challenge in many countries. To address these challenges, effective management of distribution, operation, and demand control is more necessary than ever. Leakage is one of the significant challenges in water distribution networks, and its effective management is of utmost importance for the conservation of water resources. In this context, leakage control not only reduces water loss but also lowers repair and maintenance costs. The aim of this dissertation is to develop a realistic leakage model in water distribution networks (WDNs) using field measurements and hydraulic simulation. This research aims to enhance the realism of WDN modeling by building on previous studies and addressing their weaknesses. In this regard, the development and simulation of an accurate model of WDNs using an appropriate optimization method is essential to achieve the highest level of network efficiency. In the first part of the dissertation, a practical approach is presented for estimating leakage in WDNs using calibration and available field data. In this method, pressure and flow data are initially collected, and the Density-baxsed Spatial Clustering of Applications with Noise (DBSCAN) algorithm is then applied to detect noise (outliers or abnormal data). Additionally, Bayesian optimization was used to achieve precise calibration for the network. In this section, hydraulic leak models, including the leak power relationship and the modified orifice equation, were innovatively utilized for detecting and analyzing leaks at various network nodes. The computational results indicate that this method performs with high accuracy and efficiency in the face of computational complexities and time constraints. Specifically, the calibration of a real water distribution network yielded MSE values of 0.066, 0.002, and 0.824 for the downstream pressure-reducing valve, input flow, and pressure at the critical node of the network, respectively. Furthermore, the average R² and NSE values were close to 0.98, demonstrating the model's high accuracy and alignment with real data. The leak modeling also highlighted deficiencies in the relationship between the modified orifice equation and the leak power equation. This part of the research provides valuable recommendations for improving water loss management, which is of significant importance, especially for water utilities with limited budgets and revenue. The second part of this dissertation addresses pressure management in WDNs with the aim of reducing leakage and maintaining the minimum required pressure at every part of the network through the installation of pressure-reducing valves (PRVs). In this study, the location and setting of these valves are determined using graph theory. First, the pressure at nodes and the flow rate along pipes are calculated using the standard hydraulic solver of the open-source software EPANET. Then, the hydraulic model is used as a directed graph, considering the flow direction, to identify the connections that have the greatest impact on reducing downstream pressure. Subsequently, the pipes with the most significant impact are selected for PRV installation, and the locations of the valves are determined. In the end, after each execution of the algorithm, a PRV is added to the hydraulic model, including its installation location and the proposed downstream pressure. The number of algorithm iterations depends on the number of valves required for installation. The placement of PRVs using the proposed algorithm has been implemented in two case studies. The results show that with the installation of the first PRV, a significant reduction in the overall network pressure is achieved. It is noteworthy that the pressure at the nodes does not drop below the minimum level required for service delivery. The third section highlights that calculating total network leakage does not require the pressure at the Average Zone Point (AZP). Instead, total leakage can be estimated with high accuracy using only the coefficients and nodal pressures for each section of the network.