A 28-week feeding trial was conducted in concrete tanks with Nile tilapia, Oreochromis niloticus (L.) with an average initial weight and length of 61.9±6.03 (g fish−1) and 17.6±0.45 (cm fish−1), respectively, to examine the effect of two protein levels and three feeding levels (% body weight (BW) day−1) on growth performance, production traits and body composition. Twelve 4-m3 concrete tanks (2 × 2 × 1.25 m, long, width and height) were each stocked with 100 fish and fed diets containing either 25% or 30% crude protein at rates of 1%, 2% and 3% BW daily (2 × 3 factorial experiment). The results revealed that there was no significant increase in growth rate with increasing dietary protein levels, whereas there was significant increase in growth rate with increasing feeding levels (P≤0.05). The same trend was also observed for mean BW (g), mean body length (cm), production rate (kg m−3), specific growth rate (SGR % day−1), feed conversion ratio (FCR), condition factor (K) and survival rate (%). The best final mean BW (g), final mean body length (cm), SGR (% day−1), FCR, K, production rate (kg m−3) and survival rate (%) were recorded in groups of fish fed with 25% dietary protein at the 2% feeding level. Whole fish fat and energy contents were not significantly influenced (P>0.05) by protein levels and feeding levels. Protein and ash contents were significantly (P≤0.05) influenced by feeding level, but not by dietary protein level. Economic evaluation indicated that dietary protein 25% (diet A) at the 2% BW day−1 feeding level was the most cost-effective and affordable feed strategy for farmers. We conclude that a 25% protein diet fed at 2% BW day−1 is recommended for adult Nile tilapia reared in concrete tanks.