To gain information about efficacy and safety of sunscreens, we compared the skin penetration of ultraviolet (UV) filters from two vehicles, i.e. an oil-in-water (O/W) emulsion gel and petrolatum jelly both in vitro and in vivo, as well as the corresponding pharmacological effect, i.e. the sun protection factor (SPF) in vivo. The UV filters studied were benzophenone-3 (BPH), ethylhexyl methoxycinnamate (EHM), butyl methoxydibenzoyl methane, ethylhexyl salicylate and homosalate. The human skin penetration of these five chemicals from the two vehicles was determined both in vitro using Franz cells and in vivo using a standardized tape-stripping method. The SPF of the two sunscreens was determined in vivo following the COLIPA guidelines. In vitro none of the filters permeated through the skin after 6 h of product application and very little could be found in the skin. BPH and EHM were the only UV filters found in the dermis (both after 30 min and 6 h). An effect of the vehicle could be noticed only for BPH after 30 min in the dermis and 6 h in both dermis and epidermis. In vivo, no differences in the amount of individual UV filters (in % of the applied dose) in the 15 first strips of the stratum corneum (SC) were found following 30 min of application of the formulations; however, the amount of UV filters that were retained in the SC was significantly higher (around 3 times) with the O/W emulsion gel than with the petrolatum jelly. This difference between the two vehicles was also of consequence for the SPF in vivo measured 30 min after application of the products (SPF ≅ 18 with the O/W emulsion gel compared to SPF ≅ 10 with the petrolatum jelly). By choosing the right vehicle or optimizing it, not only sunscreen products can be significantly improved in terms of pharmacological efficacy but the potential toxicological risk associated with the skin penetration of UV filters may be significantly reduced.