Abstarct: Drilling fluid loss is one of the major challenges encountered during oil and gas well drilling operations, which can lead to a reduction in circulating fluid volume, wellbore instability, formation damage, and increased operational costs. In this study, an artificial neural network was employed to predict the real-time drilling fluid loss rate using operational data and parameters recorded during drilling . The data used in this study were extracted from actual field data obtained from seven wells in southern Iraq. Data from wells 1, 2, 4, and 5 were divided into 70% for training and 30% for testing and were used for the development and initial evaluation of the model. The unseen data from wells 3, 6, and 7 were used for validation of the real-time mud loss prediction. Following data preprocessing and outlier removal, and considering the key role of rheological properties in formation pressure control, cuttings transport, wellbore stability, and hydraulic performance, as well as the limitations of conventional methods and existing predictive models, true vertical depth, mud weight, Marsh funnel viscosity, pH, hardness, solids content, chloride concentration, plastic viscosity, and yield point were used as model input parameters for estimating drilling fluid loss . In this study, the artificial neural network model achieved coefficients of determination of 0.983, 0.952, and 0.912 for drilling fluid loss prediction during the training, testing, and validation stages, respectively. The low error values obtained across all stages indicate the high accuracy, stable performance, and satisfactory generalization capability of the developed model for real-time prediction of drilling fluid loss.