Thousands of Swedish acidified lakes and streams have been regularly limed for about 30 years. Standard sampling of fish assemblages in lakes and streams was an important part of monitoring the trends after liming, i.e. sampling with multi-mesh gillnets in lakes (EN 14757) and electrofishing in streams (EN 14011). Monitoring data are nationally managed, in the National Register of Survey test-fishing and the Swedish Electrofishing Register. We evaluated long-term data from 1029 electrofishing sites in limed streams and gillnet sampling in 750 limed lakes, along with reference data from 195 stream sites and 101 lakes with no upstream liming in their catchments. The median year of first liming was 1986 for both streams and lakes. The proportion of limed stream sites with no fish clearly decreased with time, mean species richness and proportion of sites with brown trout (Salmo trutta) recruits increased. There were no consistent trends in fish occurrence or species richness at non-limed sites, but occurrence of brown trout recruits also increased in acid as well as neutral reference streams. Abundance of brown trout, perch (Perca fluviatilis) and roach (Rutilus rutilus) increased significantly more at limed sites than at non-limed reference sites sampled before and after 1986. The mean species richness did not change consistently in limed lakes, but decreased in low alkalinity reference lakes, and fish abundance decreased significantly in limed as well as in non-limed lakes.