Field experiments were carried out in 2005-2006 at two localities: Lipnik near Stargard Szczeciński (soil of quality class IVB, and of good-rye-soil-complex) and Kruszyn Krajeński near Bydgoszcz (soil of quality class V, and of weak-rye-soil-complex). The experiments were designed by split-plot method in 3 replications. Two factors were considered: irrigation (control, sprinkler irrigation) and nitrogen fertilization (0, 40, 80 and 120 kg N ha-1). Yields of non-irrigated true millet grown on the soil of quality class V were dependent on rainfall amount during the vegetation. Lower yields (1.87 t ha-1) were noted in the year characterized by lower rainfall (161 mm from May 1 to August 31). Higher yields (2.32 t ha-1) were obtained in the year with higher rainfall (229 mm). Water needs of true millet during the vegetation period (May-August), estimated as optimal rainfall according to Klatt and Press, were higher on the very light soil (quality class V) at Kruszyn Krajeński (278mm and 294 mm, respectively) than those on the light soil (quality class IVB) at Lipnik (260 and 256 mm, respectively). Highest rainfall deficit during the vegetation period of true millet – estimated as difference between optimal rainfall according to Klatt and Press for a particular month and the rainfall total in the month – was noted in July. Production results of irrigation depended on the soil complex. Sprinkler irrigation significantly increased grain yields of true millet cv. ‘Gierczyckie’ by 1.79 t ha-1 (85%) on the very light soil (Kruszyn Krajeński). Under conditions of the light soil (Lipnik), sprinkler irrigation increased grain yield by 0.36 t ha-1 (14%). Under conditions of sprinkler irrigation, increased nitrogen fertilization doses (from 0 to 120 kg N ha-1) caused a significant grain yield increase of true millet to the amount 4.42 t ha-1 and 3.44 t ha-1, on the soils of quality classes V and IVB, respectively.