The Graph Isomorphism (GI) problem is a theoretically interesting problem because it has not been proven to be in P nor to be NP-complete. Babai made a breakthrough in 2015 when announcing a quasipolynomial time algorithm for GI problem. Babai's work gives the most theoretically efficient algorithm for GI, as well as a strong evidence favoring the idea that class GI $\ne$ NP and thus P $\ne$ NP. Based on Babai's algorithm, we prove that GI can further be solved by a parallel algorithm that runs in polynomial time using a quasipolynomial number of processors. We achieve that result by identifying the bottlenecks in Babai's algorithms and parallelizing them. In particular, we prove that color refinement can be computed in parallel logarithmic time using a polynomial number of processors, and the $k$-dimensional WL refinement can be computed in parallel polynomial time using a quasipolynomial number of processors. Our work suggests that Graph Isomorphism and GI-complete problems can be computed efficiently in a parallel computer, and provides insights on speeding up parallel GI programs in practice.