Most of the world’s liquid freshwater supply is kept in groundwater, which is increasingly strained due to overdraft. Groundwater Resource (GWR) is one of major sources for agricultural crop production, water supply, land development, and economic progress. Accurate simulation and forecasting of Groundwater Levels (GWLs) will aid in the sustainable management of GWR. Main objective of present study is to simulate and forecast monthly GWLs using univariate Artificial Intelligence techniques such as M5 Model Tree (M5-MT), Naive Bayes Tree (NBT), and Radial Basis Function Support Vector Machine (RBF SVM) at three observation well locations in Gadilam River Basin of India, that are Kullanchavadi, Porto Novo, and Cuddalore. Based on observed data for each model, this study suggests a methodology for simulating the time series of GWLs. Time series analysis was used in this framework to extract time-dependent characteristics of basin from groundwater data. Next, simulation models to simulate monthly GWLs have been developed for training period (Jan 1998–Jun 2008) and testing period (Jul 2008–Dec 2012), respectively, using time-domain characteristics extracted using M5-MT, NBT, and RBF SVM, respectively. The developed models were utilized to forecast monthly groundwater levels from January 2013 to December 2017. The outputs from these three models were evaluated using statistical Performance indices such as Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), and R2. M5-MT model outperformed all other models considered that are NBT and RBF SVM models, with R2 values in testing period of 0.86, 0.84, and 0.81 for Cuddalore, Kullanchavadi, and Porto Novo, respectively. The study’s data-driven modelling approach will aid in making sustainable policy decisions by estimating groundwater levels at larger scales in areas with scant data.