In this study, an AQbD-compliant chaotropic chromatography method for ziprasidone and the determination of its five impurities was developed. The influence of critical method parameters (initial and final methanol fraction in the mobile phase, gradient duration) on the set of selected critical method attributes (t_imp. V, t_imp. V − t_imp. I, S and <WUSP>) was studied by Box–Behnken design. The errors resulting from the calculation of the model coefficients were propagated to the selected responses by Monte Carlo simulations, and their predictive distribution was obtained. The design space was computed (π ≥ 80%), and a working point was selected: initial methanol fraction 38.5%, final methanol fraction 77.5%, and gradient duration 16.25 min. Furthermore, the quantitative robustness of the developed method was tested using the Plackett–Burman design. P_imp II and P_imp V were found to be significantly affected, the first by mobile phase flow rate and the second by gradient duration. Finally, the method was validated, and its reliability for routine quality control in capsules was confirmed.