The study was conducted at the Experimental Farm Rafael Fernandes, Mossoró city, Rio Grande do Norte State, Brazil, with the objective to verify the effect of irrigation levels and planting densities on the castor bean production, BRS ENERGIA . Thus, we used the experimental design in randomized blocks in a split-plot 5 x 2, with four replications. The treatments consisted of irrigation levels (0.6, 0.8, 1.0, 1.2 and 1.4 of ETc) in main plots and plant densities in the sub-plots formed by arrangements 1.50 m x 0.30 m x 0.30 m (D1 - 37,037 plants ha-1) and 1.50 m x 0.30 m x 0.40 m (D2 - 27,778 plants ha-1). The main plots were 18 m in length, 9 m for each of the densities. The irrigation levels were differentiated by the application time using the tubes, each of which had a strip of the water distribution system independent. The main bunches fruits income decreased with the applied levels and the secondary bunches income (in fruits and grains) followed quadratic polynomial model. How much to the total income, was gotten 3,974 kg ha-1 of fruits and 2,691 kg ha-1 of grains with irrigation based on 1,20 and 1,25 of the etc, respectively.