Nematodes are keystone actors of soil, freshwater and marine ecosystems, but the complexity of morphological identification has limited broad-scale monitoring of nematode biodiversity. DNA metabarcoding is increasingly used to assess nematode diversity but requires universal primers with high taxonomic coverage and high taxonomic resolution. Several primers have been proposed for the metabarcoding of nematode diversity, many of which target the 18S rRNA gene. In silico analyses have a great potential to assess key parameters of primers, including taxonomic coverage, resolution and specificity. Based on a recently-available reference database, we tested in silico the performance of fourteen commonly used and one newly optimized primer for nematode metabarcoding. Most primers showed very good coverage, amplifying most of the sequences in the reference database, while four markers showed limited coverage. All primers showed good taxonomic resolution. Resolution was particularly good if the aim was the identification of higher-level taxa, such as genera or families. Overall, species-level resolution was higher for primers amplifying long fragments. None of the primers was highly specific for nematodes as, despite some variation, they all amplified a large number of other eukaryotes. Differences in performance across primers highlight the complexity of the choice of markers appropriate for the metabarcoding of nematodes, which depends on a trade-off between taxonomic resolution and the length of amplified fragments. Our in silico analyses provide new insights for the identification of the most appropriate primers, depending on the study goals and the origin of DNA samples. This represents an essential step to design and optimize metabarcoding studies assessing nematode diversity.