A total of 40 lambs were divided into four different treatments according to the inclusion level of the macadamia nut cake: C-control (0%), M1 (6.5%), M2 (12%) and M3 (20%). Feed was provided twice a day; animal weighing along with body condition scoring occurred within a 14-day interval. The lambs were slaughtered at the end of the performance test. Analysis of variance was performed through the Mixed procedure of the SAS, as well as linear and quadratic regression analysis. The groups presented differences between the treatments (P <0.05) for dry matter intake (DMI), ethereal extract intake (EEI), consumption in relation to live weight percentage and feed conversion ratio (FCR). The lowest mean DMI was obtained by the animals that received the M2 and differed from the C treatment, whose average was the highest. The EEI was highest for the M3 group and the FCR was also better for this group. There was a linear effect for EEI and FCR, and quadratic for crude protein intake. There was no effect for carcass characteristics, and only initial pH had a decreasing linear effect. The macadamia nut cake was effective in promoting the performance of the animals, since there was an improvement in feed conversion.