The development of high-throughput sequencing (HTS) technologies has greatly improved our capacity to identify fungi and unveil their ecological roles across a variety of ecosystems. Here we provide an overview about current best practices in metabarcoding analysis of fungal communities, from experimental design through molecular and computational analyses. By re-analysing published datasets, we find that operational taxonomic units (OTUs) outperform amplified sequence variants (ASVs) in recovering fungal diversity, which is particularly evident for long markers. Additionally, analysis of the full-length ITS region allows more accurate taxonomic placement of fungi and other eukaryotes compared with the ITS2 subregion. We conclude that metabarcoding analyses of fungi are especially promising for co-analyses with the functional metagenomic or transcriptomic data, integrating fungi in the entire microbiome, recovery of novel fungal lineages and ancient organisms as well as barcoding of old specimens including type material.