Macrobenthic secondary production is widely used to assess the trophic capacity, health, and functioning of marine and freshwater ecosystems. Annual production estimates are often calculated using empirical models and based on data collected during a single period of the year. Yet, many ecosystems show seasonal variations. Although ignoring seasonality may lead to biased and inaccurate estimates of annual secondary production, it has never been tested at the community level. Using time series of macrobenthic data collected seasonally at three temperate marine coastal soft‐bottom sites, we assessed seasonal variations in biomass of macrobenthic invertebrates at both population and community levels. We then investigated how these seasonal variations affect the accuracy of annual benthic production when assessed using an empirical model and data from a single sampling event. Significant and consistent seasonal variations in biomass at the three study sites were highlighted. Macrobenthic biomass was significantly lower in late winter and higher in summer/early fall for 18 of the 30 populations analyzed and for all three communities studied. Seasonality led to inaccurate and often biased estimates of annual secondary production at the community level when based on data from a single sampling event. Bias varied by site and sampling period, but reached ∼ 50% if biomass was sampled at its annual minimum or maximum. Since monthly sampling is rarely possible, we suggest that ecologists account for uncertainty in annual production estimates caused by seasonality.