Synthetic coagulants and flocculants used industrially in palm oil mill effluent (POME) treatment poses environmental and health risks, resulting in greater demand for natural coagulants and flocculants. This study reports the efficiency of Trigonella foenum-graecum (fenugreek) and Hibiscus esculentusas (okra) as a coagulant and flocculant respectively, for POME treatment and aims to model the process via Artificial Neural Network (ANN). The standard jar test method was used to study the effects of fenugreek dosage, okra dosage, pH and rapid mixing speed on the removal efficiency of total suspended solids (TSS), turbidity and chemical oxygen demand (COD). Optimal fenugreek and okra dosages were established at 4.6 g/l and 40 ml/500 ml POME respectively, with pH of 4.0 and rapid mixing speed of 155 rpm. The coagulation-flocculation process was modeled using a multivariable backpropagation feedforward neural network (BFNN). Model fitness was assessed using the mean squared error (MSE), mean absolute error (MAE), and mean absolute percentage error (MAPE) against 11 training algorithms with varying hidden neurons. Results indicate that Broyden–Fletcher–Goldfarb–Shanno (BFGS) training algorithm had the best performance, showing the least error values in most cases when compared to other training algorithms. The optimal hidden neuron for this the BFGS training algorithm is 17, as the MSE value for COD removal is 94.25, lowest among all training algorithms. MSE, MAE and MAPE values for TSS removal using this algorithm are 42.71, 4.76, and 0.02. MSE, MAE and MAPE values for turbidity removal are 9.49, 2.00, and 0.08, whereas MSE, MAE and MAPE values for COD removal using this algorithm are 94.25, 7.37, and 0.20. Characterization studies show that the coagulation-flocculation was influenced by inter-particle bridging mechanism. This was supported by the presence of polysaccharides found in fenugreek and okra, the interaction of sludge micro-flocs and fenugreek captured under magnification and the BET surface area of fenugreek. The elemental composition of the recovered sludge from treated POME indicates potential as a clean biomass fuel source.