In this study, Fe/Quasi MOF UIO-66 (Zr) nanocomposite was synthesized via an ultrasonic method and characterized using XRD, FTIR, SEM, and N2 physisorption analysis. The adsorption performance for methylene blue removal from aqueous solution was evaluated under varying parameters including adsorbent dosage (5-45 mg), pH (3-11), initial dye concentration (10⁻⁵-10⁻⁹ M), contact time (3-20 min), and temperature (25-80 °C). Experimental optimization revealed that maximum removal efficiency (>96%) was achieved at 45 mg adsorbent, pH 11, initial concentration 10⁻⁹ M, time 12 min, and temperature 70 °C. The core innovation of this work lies in the development of a GMDH (Group Method of Data Handling) neural network combined with a genetic algorithm to model the removal process. The optimized model predicted residual dye concentration with high accuracy (test RMSE = 0.00684, R² = 0.9922). Sensitivity analysis based on input variable frequency in the optimal chromosome revealed that adsorbent dosage (8 repeats), pH (7 repeats), and contact time (7 repeats) exerted the greatest influence on process performance. The proposed modeling approach offers a reliable analytical tool for predicting and optimizing wastewater treatment processes without extensive experimentation.