<p>Evaluating the intensity of paddy cultivation and tracking the evolving patterns of groundwater levels is essential for a comprehensive assessment of the impact of paddy farming on water resources across diverse geographic regions. Due to the notable water-intensive nature of paddy cultivation, it is imperative to take into account the condition of groundwater resources. In regions such as the cloud-prone Lower Gangetic Plain, characterized by small landholdings, traditional coarse-resolution optical remote sensing methods are inadequate in delivering sufficient information. Therefore, a comprehensive framework that integrates advanced machine learning techniques with multi-temporal SAR datasets becomes essential for assessing the spatio-temporal variations in paddy intensity. In this investigation, Sentinel SAR images of 2022 were employed to evaluate the effectiveness of a machine learning model for mapping paddy rice at a 10-m scale. The analysis of multi-temporal Sentinel-1A data revealed a trajectory of rice growth phases that aligns with the crop calendar of the study area. Notably, the characteristics of σ VH (backscattering coefficient) demonstrated a strong correlation with the growth stages of rice throughout the Aus, Aman, and Boro seasons. This study highlighted the superiority of the machine learning-based Random Forest (RF) model, which achieved an overall accuracy of ≥ 90% in identifying areas under paddy cultivation and discerning their seasonal distribution in cloud-prone regions. Regarding groundwater variability, an assessment was conducted using Central Groundwater Board (CGWB) data spanning from 2002 to 2022. The results indicated that 70% of the paddy-cultivated area in the study region exhibited a decreasing trend in groundwater levels. This integrated assessment provides a holistic understanding of the interplay between paddy cultivation intensity and its impact on groundwater resources, serving as a valuable resource for sustainable land and water management.</p>