The applications of evolutionary trees have been widely applied in many related biological and biomedical problems. These problems can be multiple sequence alignment, protein structure and its associated function predictions, and drug design. Many of them such as phylogenies construction and multiple sequence alignment are very computationally demanding. Constructing any evolutionary tree is very time consuming and has been recognized as a NP problem. In this thesis, we proposed a DNA algorithm for any evolutionary tree construction based upon bioinformatics computing. This DNA algorithm is fully utilizing parallelism to conquer algorithm complexity bottleneck constructs any evolutionary tree much more efficient. The experimental results shows that the complexity for all of different evolutionary tree constructions in the same time is in O(n3) polynomial bound.