The local co-circulation of multiple phylogenetic lineages is particularly likely for rapidly evolving pathogens in the current globalization context. When various phylogenetic lineages co-occur in the same fields, they may simultaneously be present in the same host plant (i.e. mixed infection), with potential important consequences for disease outcome. This is the case in Burkina Faso for the rice yellow mottle virus (RYMV), endemic to Africa, where it constitutes a major constraint to rice production. We aimed at deciphering the distinct RYMV isolates simultaneously infecting a single rice plant and sequencing their genomes. To this purpose, we tested various sequencing strategies, and we finally combined direct cDNA ONT (Oxford Nanopore Technology) sequencing with the bioinformatics tool RVhaplo. This methodology was validated though the successful reconstruction of two viral genomes distant from as less as a hundred nucleotides (out of 4450nt length genome, i.e. 2-3%), and present within artificial mixes at up to a 99/1 ratio. Then, we used this method to subsequently analyze mixed infections from field samples, revealing up to three RYMV isolates within one single rice plant sample from Burkina Faso. In most cases, the complete genome sequences were obtained, which is particularly important to better estimate the viral diversity and permits to detect recombination events. The described methodology consequently allows to identify various haplotypes of RYMV simultaneously infecting a single rice plant, obtain their full-length sequences, as well as a rough estimate of relative frequencies within the sample. It is efficient, cost-effective, as well as portable, so that it could further be implemented where RYMV is endemic. Perspectives include to decipher mixed infections involving other RNA viruses threatening crop production worldwide.