Secondary organic aerosol (SOA) accounts for a large fraction of the tropospheric particulate matter. Although SOA production rates and mechanisms have been extensively investigated, loss pathways remain uncertain. Most large-scale chemistry and transport models account for mechanical deposition of SOA but not chemical losses such as photolysis. There is also a paucity of laboratory measurements of SOA photolysis, which limits how well photolytic losses can be modeled. Here, we show, through a combined experimental and modeling approach, that photolytic loss of SOA mass significantly alters SOA budget predictions. Using environmental chamber experiments at variable relative humidity between 0 and 60%, we find that SOA produced from several biogenic volatile organic compounds undergoes photolysis-induced mass loss at rates between 0 and 2.2 ± 0.4% of nitrogen dioxide (NO2) photolysis, equivalent to average atmospheric lifetimes as short as 10 h. We incorporate our photolysis rates into a regional chemical transport model to test the sensitivity of predicted SOA mass concentrations to photolytic losses. The addition of photolysis causes a ∼50% reduction in biogenic SOA loadings over the Amazon, indicating that photolysis exerts a substantial control over the atmospheric SOA lifetime, with a likely dependence upon the SOA molecular composition and thus production mechanisms.