Pearl millet [Pennisetum glaucum (L.) R.Br.] production under subsistence farmer management on the sandy soils of southwestern Niger is faced with many challenges, including declining soil fertility, highly variable and scarce rainfall and poor resource base of the peasant farmers in the region. This study was conducted to evaluate the potential of management to increase yield and water use efficiency of pearl millet grown on two farmers’ fields in Niger during two growing seasons, 2003 and 2004. The management practices tested were: 1) Five manure treatments (no manure, transported manure, current corralling, a year after corralling, and two years after corralling); 2) The microdose technology (20 kg di-ammonium phosphate ha-1, and 20 kg di-ammonium phosphate ha-1 + 10 kg urea ha-1); and lastly, 3) Three different pearl millet cultivars (Heini Kirei, Zatib, and ICMV IS 89305). In both growing seasons, manure had the greatest effect on the yield and water use of pearl millet at both sites. In 2003 grain yields were 389 kg ha-1 in the NM treatment and 1495 kg ha-1 in the C0 treatment at Banizoumbou whereas at Bagoua, the NM treatment had 423 kg ha-1 vs. 995 kg ha-1 in the C0 treatment. In 2004, the NM treatment at Banizoumbou had 123 kg ha-1 grain yield and the C0 treatment had 957 kg ha-1 whereas at Bagoua the NM treatment had 506 kg ha-1 vs. 1152 kg ha-1 in the C0 treatment. Residual effects of manure led to grain yields in the C1 and C2 treatments which were more than twice as high as in the NM treatment. The improved cultivars were generally superior for grain yields, whereas the local landrace was superior for straw yields at both sites. Root zone drainage was decreased by between 50 to 100 mm, and water use increased by the same amount in the current corrals at the two sites during the two growing seasons. Increased water use under corralling and presence of residual profile moisture at the end of each of the two seasons suggested that water did not limit pearl millet production at the two sites.