Energy use in water supply systems represents a consequent part of global energy consumption across all sectors. This consumption is expected to rise, due to the increasing demand and the recourse to unconventional water resources. Regarding water utilities, most of their operating costs are related to energy consumption, especially pumping systems consumption. In this context, the main objective of this study was to model accurately by using data statistical analysis the energy consumption of pumping systems in order to optimize the whole water supply system, thus improving its efficiency, especially in the case of a limited renovation. For this purpose, Multiple Linear Regression was fitted to model the produced kWh/m 3 ratio costs according to the following Key-parameters associated to drinking pumping stations: i) active and reactive energies; ii) the daily produced water volume; iii) the power factor (Cosφ); iiii) and the operating time of each pump. The final model describes accurately the consumption per cubic meter produced with R-square statistic reaching 0.91 and value standard error is close to 5% were found. Therefore, this model could be considered a good estimator for the calculated ratio, which was close to the experimental one. In addition, this approach considers the system in the Real-Time-Data behavior, while most of the comparable studies focus on the pump scheduling problem estimator for the calculated ratio which was close to the experimental one.