Environmental DNA (eDNA) is considered to be an innovative method that can be used in the detection and identification of species present in aquatic environments, adding the advantage of describing the ichthyofauna without capturing the organisms. Out of the two possible approaches of eDNA, the metabarcoding approach aims to identify the diversity of a specific taxon and its composition using a universal DNA primer. This approach has grown considerably in recent years, although most studies focus on temperate regions, which causes a lack of clarity about the potential of the method in tropical regions since these have very distinct abiotic characteristics, as well as a higher diversity of fauna and flora. Therefore, as a way to verify the potential use of eDNA metabarcoding, this study focused on the Neotropical region, specifically the Amazon basin, a place with the highest ichthyofaunistic biodiversity in the world. As such, the goal of this study was to detect the composition of the main taxonomic groups of fish in the small stream Tarumã-Mirim, an acidic blackwater river that is located nearby Manaus city and is part of an environmental protection area near the city of Manaus, Brazil. For that, water samples of 6 L were collected along the river at 5 points (totaling 30 L). The water samples were stored in ice boxes and transported to the laboratory for filtration in fiberglass filters. Following the filtration and the extraction of eDNA, the samples were amplified in PCRs machines, using the universal primers MiFish developed by Miya et al., (2015). The next steps were: the construction of eDNA libraries, the sequencing on Illumina – Miseq platform, and a bioinformatic analysis. For the taxonomic identification of the eDNA sequences we used the sequences deposited in GenBank and the reference database developed specifically for Amazonian fish. The application and efficiency of eDNA metabarcoding for the samples, identified 42 molecular operational taxonomic units (MOTUs), which were grouped into six orders and eight families. A refined approach identified seven species and one genus. In addition, since eDNA samples were obtained along the entire river, it was possible to build a distribution map of fish species. Despite the limitation in identifying the abundance of all taxonomic groups present in the Tarumã-Mirim River, the metabarcoding approach for the eDNA was able to identify the species that have their contemporary distribution in the Amazonian rivers, as well as the detection of the three main groups of fish. Thus, the method may serve as a complement to ecological studies for the detection, identification and monitoring the Amazonian ichthyofauna. These results show that the eDNA approach needs to be optimized in tropical regions since the biotic and abiotic characteristics, which are peculiar to the tropics (acidic waters, high temperatures, etc.), may affect the adequate detection of the rich biodiversity of such sites