Synopsis The protocol described here has been developed to measure the stability of UV-B filters; a modified version is recommended for UV-A filters. It should be considered as a tool to predict the effectiveness remaining after exposure to UV-A and UV-B light. It is a simple and reliable in vitro model simulating conditions of actual use. The results show that each filter requires an appropriate choice and fine tuning of reproducible analytical conditions. While absolute values are directly influenced by uncertainties in irradiance (dosimetry), comparative measurements with respect to a known standard are very reliable.