<p>Drought is a big challenge to world water security and ecosystem resilience, as defined by long dry periods leading to water scarcity. Because of the stochastic recurrence and severe socioeconomic impacts of droughts, precise drought modelling and forecasting are required for effective water resources management. Hence, exploring efficiency of shallow and deep machine learning techniques to enhance predictive accuracy of drought forecasting models is necessary. This article investigates and compares efficiency of a shallow multilayer perceptron (MLP) model with two deep models, two-hidden layer MLP (TMLP) and Long Short-Term Memory (LSTM). To this end, long-term (1950 to 2024) grid-based monthly Standardized Precipitation Evapotranspiration Index (SPEI) datasets near to the holy city of Karbala, Iraq, were retrieved and used to train and test all the models. First, the nearby grid data sets were averaged arithmetically to represent temporal variation of drought across this data scarce city. Then, the representative dataset was separated into training and testing subsets. The optimum predictors for one-month ahead SPEI forecasting scenario were determined via mutual information between SPEI and its lagged vectors. The results showed all models produce promising accuracy. A shallow MLP having 14 hidden neurons was found slightly superior to its deep TMLP and LSTM counterparts. Thus, the use of complex deep learning models is not suggested for SPEI modelling in the studied area.</p>