The aim of this study is optimization of experimental factors in a UV-visible spectrophotometric method for simultaneous determination of dorzolamide hydrochloride and timolol maleate in eye drops. For this purpose, partial least squares-1 method was developed for resolving overlapped UV-visible spectra of these compounds. Principal component number, wavelength range (number of wavelengths) and data preprocessing applications were investigated as experimental factors. The affected experimental response in this method is PRESS (Prediction Error Sum of Squares) values. Response Surface Methodology (RSM) method was used for interpretation of relationships between these factors and PRESS values, and results were shown as equations and contour graphs. Multivariate calibration techniques which were used in the method development have some advantages such as using whole spectrum for the simultaneous analysis of the mixture component that allow to minimize the effects of noise peak, shorter analysis time, low cost and simple instrumentation without any chemical pre-treatment. This optimized method was validated and applied to a pharmaceutical preparation, eye drops, and results were compared with a literature method statistically and no significant difference was observed.