[Objectives] As one of the major groups of vertebrates, the evolutionary relationship of birds has been one of the main focuses of ornithological research. Morphological and molecular characters have been widely used to infer phylogenetic relationships in birds, but morphological characters have long been limited by homoplasy, making it difficult to draw consistent conclusions about phylogenetic relationships. However, benefiting from the molecular data and sophisticated analysis methods, we have gradually reached a consensus on the phylogenetic relationships of birds in recent years. On this basis, it is necessary to re-explore the research value of the neglected morphological data. Therefore, we aim to investigative whether there is a significant difference in the level of homoplasy of different kinds of morphological characters. We proceed to exploring the role of morphological characters with a low level of homoplasy in improving the phylogenetic signals of the molecular data and the level of support on the molecular tree. [Methods] Based on a reliable molecular phylogenetic tree by Prum published in 2015 (Prum tree) and an avian morphological dataset by Livezey & Zusi published in 2006 (Livezey & Zusi morphological data), we mapped the Livezey & Zusi morphological data to the Prum tree using PAUP* 4.0a and calculated the score for the level of homoplasy. We performed the Mann-Whitney U test to explore whether morphological characters are significantly different in their levels of homoplasy. Additionally, we performed a principal component analysis to determine if there was an interconnection between the morphological characters of homoplasy. Lastly, we added the filtered low homoplasy level morphological characters (specifically those with a CI ≥ 0.5 when mapped to the Prum tree) to Hackett’s molecular data. Then we rebuilt the tree using the maximum parsimony method and mapped the character (molecular or morphological character) to the tree (molecular or combined tree). After we got homoplasy outcomes of the mapping above, we performed the Mann-Whitney U test to compare the homoplasy outcomes. Meanwhile, we compared the bootstrap scores for each branch of the two trees (molecular and combined tree) to determine whether the inclusion of morphological data would have an impact on the level of support. [Results] We found different levels of homoplasy in morphological characters, as indicated by the consistency index of the miscellaneous characters scored significantly larger than those of osteological and myological characters (P < 0.01, Table 2), and that of cranial characters was significantly larger than those of the postcranial, body and leg characters (P < 0.05, Table 2). Nevertheless, the principal component analysis indicated that morphological characters are not easily distinguishable from each other. Additionally, adding morphological characters with low levels of homoplasy significantly increased the phylogenetic signals of the molecular data (P < 0.05, Appendix 3) and increased the level of support on branch of the molecular tree. [Conclusion] In summary, morphological characters show different levels of homoplasy, and certain groups of morphological characters can increase the phylogenetic signal of the molecular characters.