Technological advancements in the field of computing, particularly in artificial intelligence (AI), have enabled the implementation of methods aimed at reducing analysis response times, with the goal of lowering costs and enhancing the quality and safety of operations. The purpose of this study is to develop an AI-based model to predict the corrosion resistance of low-carbon steel protected by electro-deposited zinc surface layers, taking into account variations in the parameters of the galvanization process. In order to assess corrosion resistance, an analysis of the electro-deposited zinc layer thickness was performed, as there is a direct relationship between these two factors, as referenced by NBR 10476: Electro-deposited zinc coatings on iron or steel. The experiment was carried out using the Hull cell method, which involves observing regions of low current density (LCD) and high current density (HCD) on the test specimen. For the experiment, a fractional factorial design was used with seven factors and two levels. The input variables were plating time (15 and 30 mins), ZnO concentration (7 and 14 g/l), NaOH concentration (105 and 140 g/l), anode material (Fe and Zn), Purifier additive (5.5 and 11 ml), Base additive (6 and 12 ml), and Brigtner additive (0.75 and 1.5 ml). The response variable was the electro-deposited zinc layer thickness. X-ray fluorescence was used to measure the layer thickness. The results of the electro-deposited zinc layer thickness in μm for LCD and HCD regions respectively were as follows: mean 3.44 and 7.85, standard deviation 1.41 and 4.63, minimum 1.13 and 2.3, and maximum 6.19 and 19.1. Initial statistical analyses of the experiment were conducted using the F2FR package in the R programming language, and first and second-order multivariate regression models were generated. Subsequently, with the statistical validation of data using F2FR, the Orange Data Mining software was employed to create an AI-based model. The AI algorithms used were linear regression, random forest, and gradient boosting (xgboost method). The first-order and second-order multivariate regression models were compared with the best-performing AI model, xgboost, which achieved superior performance with r² = 0.95 and MSE = 0.815. In conclusion, we find that the prediction method using AI is valid and outperforms analytical regression methods. Prediction methods for electro-deposited layer thickness could become valuable tools for industrial galvanoplasty processes, assisting, for instance, in pre-adjusting input parameters to achieve desired layer thicknesses.