Abstract: To supply the evidence of identification of alpaca hair color, the difference and correlation between different melanins in alpaca hair was studied. According to the visual inspection, 22 different hair color genotypes with 3 alpacas each were selected. Alpaca fleece was taken from the back and then dissolved with different solution to get melanins. The amount of total melanins (ASM), eumelanin (EM) and pheomelanin (PM) in alpaca fleece with different hair color were measured with spectrophotometric assay, sepia as a standard for the spectrophotometric assay. Then the difference and correlation between the amount of melanins were analyzed by one-way ANOVA and linear regression curve method, respectively. The results showed that there were difference in ASM (P＜0.01) and EM (P＜0.05) and no difference in PM (P＞0.05) between alpaca fleece with different hair color, and that two models were deduced from the linear regression curve analysis: model 1: Sp.ASM＝-0.313+3.168×Sp.EM+E (R2:0.902, t=-3.732, P=0.000); model 2: Sp.EM＝0.285+0.143×Sp.ASM+E(R2:0.902,t=6.769,P=0.000). It was suggested that ASM and EM could be used as the parameters to differentiate hair color of alpaca because of significant correlation to hair color, moreover, both of them have the relation of linear regression curve.