This study aims to improve accuracy and friendly techniques to predict TL by considering more input parameters and leveraging ANN. A dataset of around 309 data points from 16 previous experimental studies from the literature is used to develop the ANN model to modify the traditional TWC model accordingly. A sensitivity analysis is conducted to find the most effective parameters to modify the TWC model out of the 12 input parameters. Only five modification factors are employed: the coating condition, end condition, release condition, corrosion loss percent, and vertical spacing between strands. The developed neural network and modified TWC model show superior performance compared to existing models, with coefficients of variance of approximately 12.3% and 16.3%, respectively. The results highlight the significance of parameter inclusion in predictive models and suggest further exploration of additional parameters for enhanced accuracy in future research.