In practice, in addition to the fixed effects, the experimental design often involves one or several random effects to form a mixed effects model. Variance components of the model can be obtained from the ANOVA table by the method of moments (MME). However, MME doesn’t have the statistical property of the maximum likelihood estimator. The restricted maximum likelihood (REML) is another useful technique for analysing the mixed effects models. In this paper, we illustrate how to use REML to analyze the variance components, and compare REML with other methods.