Net sampling by trawling and hydroacoustics was used to methodologically compare size spectra (SS) of the pelagic fish community in a deep lake across 12 years of sampling. Hydroacoustic SS were generated based on either single-echo detections (SEDs) or tracked-echo groups (TEGs) from 20 cross-lake transects. Trawl SS were obtained by a midwater trawl in four pelagic depth layers. All SS were derived from maximum likelihood estimations of exponent b of a continuous fish body mass distribution. The arithmetic mean exponent b was similar for all methods, and there were no significant differences of b among the three methods across years. However, visual inspection indicated that the SS differed considerably between trawling and hydroacoustics in some of the years, primarily when high densities of 0+ coregonid fishes were strongly spatially aggregated and hence caught by the trawl. Accordingly, there was no correlation between SS generated by trawling and hydroacoustics. In contrast, SS generated by SEDs and TEGs were significantly correlated, indicating reliability and reproducibility of obtaining SS by hydroacoustics. The SS estimated by TEGs revealed a positive trend of exponent b over the years since 2005, potentially reflecting the recent eutrophication of Lake Stechlin, which may lead to higher fish growth rates. We conclude that hydroacoustics may help to generate more precise SS of the pelagic fish community in our study lake than midwater trawling. However, the truthfulness of SS estimates cannot be evaluated because of the inherent difficulty in determining the true densities and sizes of fishes in lakes.