The purpose of this study was to evaluate the impact of substrate roughness and of product spreading method on in vitro sun protection factor (SPF) measurement and to define the experimental conditions most appropriate to reach the best level of correlation to in vivo SPF. In vitro SPF assessment was carried out on 13 products (including different formulation types with SPF from 20 to 75) using various in vitro SPF protocols and comparing related predictive potential regarding in vivo SPF. In the first part, two spreading methods were compared on two types of PMMA (Polymethyl methacrylate plate with different roughness. The impact of a second spreading step after product drying was also evaluated. From the various investigated parameters, it was shown that (i) a higher roughness (Ra = 4, 5 microm) was preferred for O/W formulations (ii) using a defined sequence of light linear and circular strokes was more adequate than monitoring product spreading in terms of time and pressure (iii) both correlation to in vivo SPF and results variability were improved when a second spreading step was added. The altered protocol showed a good predictive potential regarding in vivo SPF values for O/W formulations (correlation coefficient 0.92, correlation curve slope 0.98) and coefficient of variation of in vitro results (14% of the mean SPF value) close to what is usually obtained in vivo. The repeatability of the protocol was also demonstrated. In the second part, we evaluated the impact of PMMA plate pre-treatment with paraffinum liquidum before spreading the product to get a better correlation between in vivo and in vitro SPF values for W/O formulations. This allowed us to define a protocol suitable for both O/W and W/O formulations.