One of the ways to study complex networks is making them visualized. With the advent of Web2.0 era and the big data era, the scale of the complex networks is becoming larger. That brings the new challenge for the visualization of complex networks on the layout effect and speed. This paper aims at improving the FDA(Force-Directed Algorithm) from this two aspects. On the layout effect, the articulation points of the complex networks are used to compact the networks, which achieves showing the layout result hierarchically. On the speed of the algorithm, repulsive force and attraction are computed on GPU, and the operation of updating the coordinates of vertex is also implemented on GPU. That improves the efficiency of the algorithm greatly.