In this paper, we propose an algorithm of Path-aware Siamese Graph neural network(PSG) for link prediction tasks. Firstly, PSG can capture both nodes and edge features for given two nodes, namely the structure information of k-neighborhoods and relay paths information of the nodes. Furthermore, siamese graph neural network is utilized by PSG for representation learning of two contrastive links, which are a positive link and a negative link. We evaluate the proposed algorithm PSG on a link property prediction dataset of Open Graph Benchmark (OGB), ogbl-ddi. PSG achieves top 1 performance on ogbl-ddi. The experimental results verify the superiority of PSG.