Hematite-cellulose acetate composite membranes (M1, M2, and M3) were evaluated for their performance in methylene blue (MB) adsorption. Box-Behnken design was used to evaluate the influence of operational parameters on the adsorption of MB by these adsorbents. The idea was to optimize the process with maximum efficiency on adsorption, through identification of the influential factors to the process, evaluation of interactions between these factors, and modeling mathematical expressions. Adsorption experiments versus time were better described by the non-linear pseudo-first-order model, followed by the pseudo-second-order model, and then the non-linear pseudo-first-order model. The R2 and chi-square values corroborate this finding. Equilibrium data was also evaluated and modeled to the linear and non-linear Langmuir and Freundlich isotherm models. The linear isotherm models were better models to fit the adsorption equilibrium data on the membranes, followed by the non-linear Langmuir model. Thus, it was presumed that chemisorption was the prevailing adsorption mechanism. Additionally, adsorption proved to be exothermic and spontaneous. A quadratic model equation was found to describe the interaction between factors. The coefficients of pH and MB concentration were positive, thus positively affecting the adsorption of MB, but those of the temperature were negative, justifying the negative effect on the adsorption process. Analysis of variance (ANOVA) showed a high coefficient of determination value and a high prediction of the regression model was derived. The highest adsorption capacity was found at the optimum experimental conditions of pH = 9.0, MB initial concentration = 100 mg · L-1 and temperature = 20 ℃. Finally, the composites provided 5 use cycles which is good when considering the adsorbents used in removing textile dyes from wastewater.