This study was conducted to compare five growth functions (Brody [B), Logistic [L], Gompertz [G], Von Bertalanffy [VB] and Richards [R]) in describing the body weight changes across age in three broiler ecotypes. Each chick was wing-tagged at day old and weighed on a weekly basis up to 10 weeks of age. Aforementioned non-linear growth models were fitted to appraise age-body weight relationship using procedure NLIN of the S.A.S (Version 9.1). G, L and VB functions converged with a low number of iterations ranging from 8 to 33 in Marshal, 7 to 10 in Naked Neck and 5 to 9 in Normal Feathered chickens. VB had the highest number of iterations (33, 10 and 9) for Marshal, Naked Neck and Normal Feathered chickens, respectively. The G, VB and L models fitted the growth curves of all the chicken ecotypes very well, and the fitting degrees R2 were all above 99.89. Based on all the criteria used for comparing these models in the three ecotypes, it can be established that the L function gave the best fit for the age-body relationship although G and VB functions were equally good in predicting the growth curves of the chickens. B and R functions were not good in fitting chicken growth data in this study with respect to parameter estimates, convergence criteria and p values.