It is widely accepted that evolutionary related species tend to resemble each other; therefore the traits of related species are not statistically independent. Phylogenetic comparative methods (PCMs), which incorporate phylogenetic information into statistical analysis, have been recommended in cross-taxonomic comparisons since 1980s. In PCMs, the preliminary step is to transform the raw trait data, which are usually statistically dependent, into standardized data, which are independent of the phylogeny. After this transformation, traditional statistical methods can be used for further analyses of the relationship between these traits. Over the past two decades, several PCMs such as Felsenstein′s independent contrasts technique, nested analysis of variance, and autocorrelation have been put forward and applied widely in comparative studies. In the present paper, we first introduced two common procedures before conducting PCM, phylogeny construction and traits-data diagnosis. Then, the principles and applications of PCMs were reviewed focusing on simplified independent contrasts, Felsenstein′s (1985) independent contrasts, and autoregression. Some applications of the above three methods were presented to show how efficiency they are in revealing many major patterns. However, PCMs have both theoretical and practical limitations, mainly due to the accuracy of phylogeny trees and the choosing of proper evolution models. Despite that, with the development of bioinformatics and biological systematics, as well as the exploitation of the software, PCMs is going to be widely used in comparative studies in ecology and evolutionary biology.