A method of determining the thermal properties of optical thin-film filter materials from spectrophotometric measurements is proposed. Its feasibility is shown by extracting the material properties from simulated transmission spectra imprinted with real-world measurement errors. A Monte Carlo-based algorithm was implemented to optimize thin-film filter designs for thermal stability. This algorithm is used to explore the optimization possibilities of exemplary designs for differently behaving material combinations. It is demonstrated that designs can be optimized for use under different temperatures, where also the overall wavelength shift of a filter is modifiable.