A terephthalic acid-modified chitosan−magnetic nanocomposite (Cs− Tp@Fe 3 O 4 ) was synthesized and characterized. The synthesized Cs−Tp@Fe 3 O 4 was used in a batch process for the adsorptive removal of the acid blue 25 (AB-25) dye in aqueous solutions. The kinetic data were subjected to the pseudo-first-order, pseudosecond-order, Elovich, and intra-particle diffusion models, while the equilibrium data were evaluated with the Langmuir, Freundlich, Temkin, and Dubinin−Radushkevich isotherm models. The effects of the initial dye concentration, contact time, and adsorbent dosage, as well as their interactions, on the removal efficiency were investigated using the design of experiments based on a central composite design, and the resultant data were modeled with the response surface methodology (RSM), artificial neural network (ANN), adaptive neuro-fuzzy inference system (ANFIS), and multiple linear regression (MLR) approaches. The adsorption process followed pseudo-first-order with good agreement between the experimental Q e(exp) and calculated Q e(cal.) amounts of dye adsorbed, as well as the values of correlation coefficient, R 2 (0.999) and percentage of sum square error, % SSE (0.640). All the investigated adsorption isotherms fitted all models well in the order of Dubinin−Radushkevich > Langmuir > Freundlich > Temkin with R 2 > 0.9 with the monolayer maximum adsorption capacity of 440.24 mg/g obtained from the Langmuir isotherm. The RSM model predicted the maximum removal efficiency at an optimum initial dye concentration of 19.11 mg/L, a contact time of 95.3 min, and an adsorbent dosage of 0.18 g. Statistically, the models were fitted in the order of RSM > ANN > ANFIS > MLR. These results indicated that the prepared Cs−Tp@Fe 3 O 4 is an efficient adsorbent for the AB-25 dye removal with excellent stability for water treatment applications.