The current study aims to determine how well an adsorption approach in a laboratory-scale reactor removes zinc (II) from an aqueous solution. In this work, ANFIS and the ANN were used to predict the green adsorbents' adsorption capacity in removing zinc (II). Four operational variables were studied: the contact time in minutes, the dosage of the green adsorbent in mg/100 mL, the initial concentration of the Zinc (II) solution in mg/L, and pH. The outputs were the removal percentage (%) and adsorption capacity (mg/g). The ANN and ANFIS approaches were built using 60% of the data for training, with the remaining 40% being used for validation and testing. According to the findings, the adaptive neuro fuzzy interference system models are a potential method for making predictions about the adsorption of Zinc (II). Based on this result, the training dataset's (RMSE) root-mean-square error, (AARE) Average Absolute Relative Error, (ARE) Absolute Relative Error, (MSE) Mean-Squared Error, and R² for the ANFIS model were found to be 0.021, 0.048, 0.012, 0.015, and 0.988, respectively. For the ANN model, the AARE, RMSE, ARE, MSE, and R² were found to be 0.013, 0.020, 0.021, 0.032, and 0.997, respectively. The Langmuir model best fitted the adsorption, whereas the pseudo-second order model governed the adsorption mechanism. The green adsorbent was studied for morphology, functional groups, thermal properties, and crystallinity. Although the functional groups of the adsorbent were similar to the CNCs, the TGA analysis showed that the adsorbent exhibited more excellent heat stability. The nanocomposites' needle-like shape, tiny particle size, and porous structure were all visible in the SEM.