The prediction of fluid flow rates is of prime importance for the oil and gas business. Several classical flow rate prediction correlations have been generated and generally used for wells flowing naturally. However, for oil wells flowing artificially, these correlations overlooked many significant fluid and well parameters, leading to erroneous results and inaccurate predictions. This paper used the outputs of a new mathematical model to predict the flow rates of 34 electrical submersible pump (ESP) oil wells from 3 oil fields located in North Africa, possessing a broad range of parameters, to determine a new correlation to predict the oil flow rate. Mathematical procedures for finding the best-fitting curve were used. When compared to the experimentally measured results of another two datasets, the predicted oil flow rates by the new correlation were in good agreement. Indeed, the performance of the new correlation shows reliable prediction results with about ± 10 relative error. Furthermore, the results of this new method were statistically superior to results from previous correlation methods. The correlation that was proposed here exhibits greater accuracy (�5 error) than existing previous correlation methods. The new correlation adapts itself to any ESP oil well and could apply to both onshore and offshore wells. © 2018 Elsevier B.V.