Agricultural managed aquifer recharge (Ag-MAR) techniques involve intentional enhancement of groundwater recharge through agricultural practices to increase water availability in aquifers which can be used to support irrigation. Successful implementation of Ag-MAR requires thorough hydrogeological understanding and can be supported by mathematical models to quantify infiltrated volumes and to evaluate aquifer contamination risks. Ag-MAR could be relevant for long-term sustainability of rice production in the Piedmont–Lombardy regions, where 92% of the national rice production is concentrated. Our study focuses on the Lomellina area (1250 km2), a sub-region of the Piedmont–Lombardy rice basin, bordered by the Sesia, Po and Ticino rivers. This region is an alluvial plain with flat topography (elevations between 52–133 m asl), where the phreatic aquifer is very shallow and crucial for rice production. Recently, groundwater levels are declining due to changes in climate and rice irrigation management, and the question is whether Ag-MAR can reverse this trend. In this study, to estimate the recharge from the agricultural area and the extensive irrigation channel network, a modeling framework based on a semi-distributed application of the SWAP model (https://www.swap.alterra.nl/) was implemented in QGIS and completed with a simple model simulating the percolation from the irrigation network. Then, MODFLOW-6 was used to develop a transient model which includes surface recharge and the main natural and artificial streams and canals exchanging water with the aquifer. The groundwater flow model was calibrated using groundwater level time series from 22 monitoring wells in the study area. This study was carried out in the context of the PROMEDRICE project (https://promedrice.org/; PRIMA-Section2–2022) funded, for the Italian partners, by MUR (Italian Ministry of University and Research).