Depletion of groundwater reserves has been designated as a serious problem all around the globe. Owing to rapid population growth, industrialization, and enhancement of living standards, the demand and exploitation of groundwater are rising at a rate larger than ever before. For instance, an estimate shows that the groundwater reserve in the north-western part of India depletes at a mean rate of 4.0 ± 1.0 cm/year, equivalent to a volume of 17.7 ± 4.0 km3/year. Artificial groundwater recharge is a set of man-made techniques to replenish and augment the groundwater reserve at places where the natural recharge is not adequate. Artificial recharge structures (ARSs) are used to maintain a great balance between supply and demand. The primary goal of this study is to use the analytical hierarchy process (AHP) and feed-forward artificial neural network (ANN) to map locations appropriate for ARSs. Various factors, such as elevation, slope, drainage density, distance from the river, soil, land use and land cover, geomorphology, the thickness of shallow permeable layers, population distribution, and rainfall, are used to achieve the primary objective. The secondary goal is to map the subsurface using electrical resistivity tomography in order to detect an aquifer's depth range and overall geometry. To achieve the secondary goal, the apparent resistivity (Ohm-m) is calculated by using the Wenner–Schlumberger configuration, and the true resistivity (Ohm-m) of the subsurface is then calculated by using an inversion algorithm which solves Poisson’s equation combined with Ohm’s law using the Resistivity-2D Inversion (Res2DInv) software. The tertiary goal is to suggest an appropriate type of ARS by considering a number of variables, including the availability of water, land, cost, soil, lithology, geology, and geomorphology, as well as climate, topography, and land use. The sigmoid function is considered a transfer function inside the hidden layers of a forward-based ANN, and the weights are established by an iterative process so that the error between the observed and simulated data is tolerable. AHP relies on subjective expertise, whereas ANN is driven by data, producing superior results than AHP. Based on the analysis, we advise employing surface spreading techniques for users close to the Sutlej River. Injection well-type ARSs can be advised to the locations distant from rivers.