When conducting experiments, it is often found that there are errors in the observed responses. It can cause data outliers to appear whose existence results in making conclusions inaccurate. Therefore, outliers need to be overcome by applying the robust regression method. The robust method used is the robust MM because it has a high level of efficiency and breakdown point. The Robust MM method is useful for obtaining parameter estimates in a three-factor Completely Randomized Design (CRD) which is applied to the data on average abdominal fat of broiler chickens experiencing outliers in four observations. The results showed that the presence of outliers caused no effect of differences in age of chicken and the interaction between age of chicken and feeding fermented kiambang on the average abdominal fat of broiler chickens. However, after the data was replaced with estimated data obtained from the Robust MM method to overcome outliers, it showed that there was an effect of age of chicken and the interaction between age of chicken and feeding of fermented kiambang on the average abdominal fat of broiler chickens.