Machine Learning has applications in various disciplines of water resource engineering. Various unique models are being developed these days for the prediction and study of both groundwater quality and quantity. The current work intends to predict the quantity of groundwater using the ML model Support Vector Regression, which is commonly used for regression analysis and can predict discrete values. For assessing and determining the prediction accuracy of the SVR model, 17 groundwater stations located in agroclimatic zones of Cuttack District based on agricultural district divisions of Odisha State were evaluated. Rainfall, temperature, and relative humidity were used as input parameters, with groundwater level serving as an output parameter. The outcome revealed that SVR effectively predicts Groundwater level with an accuracy of 82.35% for all stations and minimum RMSE value of 0.03217 and maximum R value of 0.8 for the station Tangi1 for the considered study region. The overall result showed that SVR is an effective ML technique for predicting the Groundwater level for the regions having agroclimatic conditions when the time-series variation of water level is very less.