Diatoms are a species-diverse phylum of microalgae often presenting high biomass in aquatic habitats. This makes them excellent ecological indicators in rivers and lakes. They are routinely used to assess ecological quality of rivers and lakes using microscopy, which is time consuming. An alternative is to determine species in samples based on short DNA barcodes and high-throughput sequencing (HTS). Former studies showed that community structure and water quality assessments based on diatoms deliver similar results with both methods. But, none evaluated if diversities were assessed in the same way despite the importance of this ecological metric. Based on littoral benthic samplings carried out in 56 pristine alpine lakes, we compared different diversity indices measured with microscopy and metabarcoding. Each lake was sampled in three different places of its littoral. We showed that α (diversity measured in a single sampling site of a given lake) and ϒ (total diversity in a lake where three independent samples were considered) diversities obtained with HTS were higher than those obtained with microscopy. This may be explained by the capacity of HTS to detect morphologically cryptic species and to better detect rare taxa. On the other hand, β diversity obtained with HTS was smaller, which may be explained by the capacity of HTS to detect very rare species and free-floating extracellular DNA. Nevertheless, diversity indices obtained with both methodologies were well correlated each other. This study validates the possibility to assess diatom diversity with HTS in a comparable way to the classical microscopic analysis.