<p>Machine learning (ML) techniques have shown significant potential in individual optimization of electrocoagulation (EC) and electrooxidation (EO) processes, but their application to combined EC and EO processes represents a significant research gap. The hybrid EC-EO process is recognized as a highly complex electrochemical process, exhibiting nonlinear relationships between operating parameters. Hence, this study investigates the implementation of data-driven models comprising artificial neural networks (ANN), adaptive neuro-fuzzy inference system (ANFIS), and support vector regression (SVR) to predict chemical oxygen demand (COD) removal efficiency from real domestic sewage water in hybrid EC-EO process. Five input variables were employed: initial pH, current intensity, electrocoagulation time, electrooxidation time, and electrode type. The ANN performance was measured using Levenberg–Marquardt (trainlm) algorithm, whereas ANFIS modelling was accomplished using Takagi–Sugeno type FIS. For SVR, an advanced Bayesian algorithm was adopted to optimize the model. The experimental datasets were segregated into 70% for training and 30% for testing. The statistical metrics, Coefficient of determination (R²), mean absolute error (MAE), mean squared error (MSE), root mean squared error (RMSE), and sum of squared error (SSE) were utilized to compare the efficiency of model with measured COD removal values. The comparative evaluation revealed model performance in the following order: ANN &gt; ANFIS &gt; SVR. During testing, the R², MAE, MSE, RMSE, and SSE values for ANN obtained were 0.9157, 0.0736, 0.0082, 0.0906, and 0.4265, respectively. Additionally, sensitivity analysis utilizing Garson’s algorithm and Partial derivatives was conducted on the superior model, establishing current intensity and electrode type as the most impactful parameters affecting COD removal efficiency in hybrid EC-EO process. This study confirms the effective implementation of data-driven techniques in hybrid electrochemical process.</p>