In this study we investigated the phosphorus (P) content of Baltic soils. In the first set of analyses, 99 soil samples from the Baltic states and Sweden (soil set 1, representing seven different catchments or experimental plots) were analysed for soil P using four extraction methods: ammonium lactate (PAL), double lactate (PDL), Mehlich 3 (PM3) and carbonate (POlsen) (r = 0.85–0.97). In absolute values, results from PM3, PDL and POlsen gave means of 71, 61 and 20%, respectively, of the value from the PAL extraction method. Significantly different relationships were found between P soil concentrations and pH of the extract. In addition, soil pH and organic matter content were found to be of importance. Secondly, we tested 110 soil samples (soil set 2) from five different Swedish monitoring fields with clay soils where PAL clearly correlated with soil P extracted in calcium chloride (PCaCl2) (r = 0.95). Values of a single-point phosphorus sorption index (PSI) correlated with the aluminium concentration (AlAL) in the lactate extract (r = 0.91) and with (AlOX) in the oxalate extract (r = 0.96). None of the soil P tests with different extraction agents – calcium chloride (PCaCl2), water (Pw), POlsen or PAL– correlated with the mean annual flow-weighted concentration (1999–2010) of dissolved reactive P (DRP) in drainage water. Neither was there any clear relationship between DRP concentration in drainage water and these tests combined with PSI or with other sorption indices including extracted Al and iron (Fe). However, DRP was related to the clay content of the topsoil (r = 0.91, P < 0.05).