This work uses three-dimensional green and biodegradable adsorbent from cellulose nanocrystals and a machine learning technique to simulate and optimize the removal of zinc (II) from synthetic acid mine drainage. The adsorption process was modelled and optimized using three machine learning algorithms: Response Surface Methodology (RSM), Adaptive Neuro-Fuzzy Inference System (ANFIS), and Artificial Neural Network (ANN). The predictive modelling of the RSM, ANN and ANFIS models demonstrated good correlation with R² of 0.987, 0.992 and 0.997 respectively. According to the findings, the created models successfully predicted the adsorption behaviour, with the AMFIS model performing best with the lowest error rate. Low values of calculated error functions of MPSD (ANFIS=0.0108; RSM=0.0199 and ANN=0.0122), RMSE (ANFIS=0.0015; RSM=0.0218 and ANN=0.01328), ARE (ANFIS=0.0011; RSM=0.0118 and ANN=0.0153, and HYBRID (ANFIS=0.0005; RSM=0.0331 and ANN=0.0098) indicated good harmony between experimental values and models’ predictions. The result showed that the order of the models’ effectiveness for Zinc (II) removal is: ANFIS > ANN > RSM. RSM was used to optimize the process, and the ideal conditions for maximal Zinc (II) removal efficiency were established. Initial pH of 6, contact time of 300 min, initial concentration of 250 mg/L, and sorbent dose of 15 mg and adsorption capacity of 350.23 mg/g was the optimal condition. The isotherm investigation demonstrated that the Freundlich isotherm with R² of 0.995 best represented the equilibrium modelling, however the kinetic analysis revealed that the pseudo second order (R² = 0.998) and Elovich (R² = 0.995) models better accounted for the kinetics of the experimental data. The study's findings might help develop cost-effective and efficient systems for treating polluted water supplies.