Although the treatment of eutrophied lakes with aluminium (Al) compounds has been established for more than 40 years, publications reporting on long-term consequences for phytoplankton are rare. Here we present observations from Lake Barleber for the period 1985–2016. The lake was treated in autumn 1986 because of high phosphorus (P) concentrations and cyanobacteria blooms, which limited the lake's recreational use. Within six weeks 480 t of Al sulphate solution (37 t of Al) were applied to the lake. This was equivalent to a dose of 36 g Al3+ m−2 or 5.7 mg Al3+ L−1. Already after having applied half of the Al sulphate, the concentration of soluble reactive phosphorus (SRP) reached its analytical limit of quantification (3 μg L−1). Removal rates calculated after completion of the treatment were 98% for SRP and 90% for total phosphorus (TP). In the following 13 years from 1987 to 1999, cyanobacteria were almost absent. In the years 2000 to 2003 as well as in 2005 and 2014 they appeared in low abundances. In the period 1987–2014, almost complete absence of cyanobacteria and high transparency provided good conditions for recreational use of Lake Barleber. Compared to pre-treatment conditions, phytoplankton biomass increased temporarily from 1987 to 2016. This increase in biomass did not interfere with the use for bathing and swimming, because phytoplankton community composition changed towards a dominance of chlorophytes and dinophytes. In 2016, however, P concentration and cyanobacterial biomass rose again to the level of the last pre-treatment years (TP 134 μg L−1, cyanobacterial biomass 1 mg L−1; averages for the period May–October). We conclude that Al treatment is effective and can last for decades. For recreational lake use, the effects of the alum treatment on phytoplankton community composition showed to be more important than its effects on total phytoplankton biomass.