In Bautzen Reservoir (Germany), a technique of internally dosing iron compounds in combination with a local water column destratification was developed in order to control Microcystis blooms. In this paper, experimental results concerning the phosphorus (P) precipitation by iron application are reported. First, preliminary studies were conducted serving the choice of an appropriate precipitant. Subsequently, a whole-lake experiment was carried out. Fe 2+ salts displayed a delayed oxidation and flocculation behaviour when injected into an ascending water jet. Nevertheless, the efficiency of soluble reactive phosphorus (SRP) elimination was equal to experiments with Fe 3+ compounds both in laboratory and field experiments. During Fe 2+ application, a large P-binding potential remained during the horizontal spreading of the treated water, since the proportion of the dissolved iron fraction near the dosage point was still high. Thus, measurements at a greater distance from the dosage point revealed higher amounts of P eliminated by Fe 2+ than by Fe 3+ dosage. Compared to the preceding year 1995, during the treatment periods in May–August 1996 and 1997 the SRP contents in the whole water body dropped by 72% and 54%, respectively, while the total phosphorus contents dropped by 45% at each period. We conclude that in-lake dosage of Fe 2+ is an appropriate method to reduce the P loading of hypereutrophic reservoirs.