Candida species are the most common cause of fungal infections around the world, associated with superficial and even deepseated infections. In clinical practice, there is great significance in identifying different Candida species because of their respective characteristics. However, current technologies have difficulty in onsite species identification due to long turnover time, high cost of reagents and instruments, or limited detection performance. We developed a seminested recombinase polymerase amplification (RPA) genoarray as well as an integrated system for highly specific identification of four Candida species with a simple design of primers, high detection sensitivity, fast turnover time, and good cost-effectiveness. The system constructed to perform the assay consists of a rapid sample processing module for nucleic acid release from fungal samples in 15 min and a digital microfluidic platform for precise and efficient detection reactions in 35 min. Therefore, our system could automatically identify specific Candida species, with a reagent consumption of only 2.5 μL of the RPA reaction mixture per target and no cross-reaction. Its detection sensitivity for four Candida species achieved 10 1 −10 2 CFU/mL, which was 10-fold better than conventional RPA and even comparable to a common polymerase chain reaction. Evaluated by using cultured samples and 24 clinical samples, our system shows great applicability to onsite multiplex nucleic acid analysis.