Predicting friction factors for wire-wrapped fuel assemblies relies on existing empirical correlations, which lack universality and show notable errors for the China Experimental Fast Reactor (CEFR). This study aims to develop a reliable and generally predictive model and introduce a novel friction factor correlation that captures the complex relationships among fuel assembly parameters, fluid properties, and friction factors. The proposed correlation can enhance the accuracy of friction factor predictions for CEFR, addressing current limitations. Several experimental datasets were analyzed using an artificial neural network (ANN) with random forest weighting to model nonlinear dependencies. Symbolic regression was employed to derive explicit mathematical expressions for the friction factor correlation as a function of five key parameters. The present study demonstrates that the ANN model demonstrated high predictive accuracy with minimal errors, while symbolic regression provided valuable insights into the fluid dynamics of wire-wrapped fuel assemblies. By integrating ANN and symbolic regression, this study offers an innovative framework for tackling complex engineering challenges, contributing to the improved design and safety analysis of nuclear reactors, especially for CEFR.