Although the number of studies investigating mycorrhizal associations in orchids has increased in recent years, the fungal communities associating with orchids and how they differ between species and sites remain unclear. Recent research has indicated that individual orchid plants may associate with several fungi concurrently, implying that to study mycorrhizal associations in orchids the fungal community should be assessed, rather than the presence of individual dominant fungal species or strains. High-throughput sequencing methods, such as 454 pyrosequencing, are increasingly used as the primary tool for such analyses. However, many studies combine universal primers from previous phylogenetic or ecological studies to generate amplicons suitable for 454 pyrosequencing without first critically evaluating their performance, potentially resulting in biased fungal community descriptions. Here, following in silico primer analysis we evaluated the performance of different combinations of existing PCR primers to characterize orchid mycorrhizal communities using 454 pyrosequencing by analysis of both an artificially assembled community of mycorrhizal fungi isolated from diverse orchid species and root samples from three different orchid species (Anacamptis morio, Ophrys tenthredinifera and Serapias lingua). Our results indicate that primer pairs ITS3/ITS4OF and ITS86F/ITS4, targeting the internal transcribed spacer-2 (ITS-2) region, outperformed other tested primer pairs in terms of number of reads, number of operational taxonomic units recovered from the artificial community and number of different orchid mycorrhizal associating families detected in the orchid samples. Additionally, we show the complementary specificity of both primer pairs, making them highly suitable for tandem use when studying the diversity of orchid mycorrhizal communities.