Conservation biology must set geographic conservation priorities not only based on the compositional or structural but also on the functional dimensions of biodiversity. However, assessing functional diversity is challenging at the regional scale. We propose the use of satellite-derived Ecosystem Functional Types (EFTs), defined here as patches of land surface that share similar primary production dynamics, to incorporate such aspects of ecosystem functional diversity into the selection of protected areas. We applied the EFT approach to the Baja California Peninsula, Mexico, to characterize the regional heterogeneity of primary production dynamics in terms of EFTs; to set conservation priorities based on EFT richness and rarity; and to explore whether such EFT-based conservation priorities were consistent with and/or complementary to previous assessments focused on biodiversity composition and structure. EFTs were identified based on three ecosystem functional attributes derived from seasonal dynamics of the Enhanced Vegetation Index: the annual mean (proxy of primary production), the seasonal coefficient of variation (descriptor of seasonality), and the date of maximum (indicator of phenology). EFT-based priorities identified 26% of the peninsula as being of extreme or high priority and reinforced the value of the ecosystem functional diversity of areas already prioritized by traditional conservation assessments. In addition, our study revealed that biodiversity composition- and structure-based assessments had not identified the full range of important areas for EFT diversity and tended to better capture areas of high EFT rarity than those of high EFT richness. Our EFT-based assessment demonstrates how remotely sensed regional heterogeneity in ecosystem functions could reinforce and complement traditional conservation priority setting.