BACKGROUND: Greater Mekong Subregion countries are committed to eliminating Plasmodium falciparum malaria by 2025. Current elimination interventions target infections at parasite densities that can be detected by standard microscopy or rapid diagnostic tests (RDTs). More sensitive detection methods have been developed to detect lower density "asymptomatic" infections that may represent an important transmission reservoir. These ultrasensitive polymerase chain reaction (usPCR) tests have been used to identify target populations for mass drug administration (MDA). To date, malaria usPCR tests have used either venous or capillary blood sampling, which entails complex sample collection, processing and shipping requirements. An ultrasensitive method performed on standard dried blood spots (DBS) would greatly facilitate the molecular surveillance studies needed for targeting elimination interventions. METHODS: A highly sensitive method for detecting Plasmodium falciparum and P. vivax 18S ribosomal RNA from DBS was developed by empirically optimizing nucleic acid extraction conditions. The limit of detection (LoD) was determined using spiked DBS samples that were dried and stored under simulated field conditions. Further, to assess its utility for routine molecular surveillance, two cross-sectional surveys were performed in Myanmar during the wet and dry seasons. RESULTS: The lower LoD of the DBS-based ultrasensitive assay was 20 parasites/mL for DBS collected on Whatman 3MM filter paper and 23 parasites/mL for Whatman 903 Protein Saver cards-equivalent to 1 parasite per 50 µL DBS. This is about 5000-fold more sensitive than standard RDTs and similar to the LoD of ≤16-22 parasites/mL reported for other ultrasensitive methods based on whole blood. In two cross-sectional surveys in Myanmar, nearly identical prevalence estimates were obtained from contemporaneous DBS samples and capillary blood samples collected during the wet and dry season. CONCLUSIONS: The DBS-based ultrasensitive method described in this study shows equal sensitivity as previously described methods based on whole blood, both in its limit of detection and prevalence estimates in two field surveys. The reduced cost and complexity of this method will allow for the scale-up of surveillance studies to target MDA and other malaria elimination interventions, and help lead to a better understanding of the epidemiology of low-density malaria infections.