Abstract: Choosing suitable and validate attack strategies has meaningful significance to smash target networks or collapse criminal groups. To predict the effectiveness of attack strategies, two new measure indexes are proposed in this research: betweenness-centrality compactness index and closeness-centrality compactness index. The relationship of the indexes of average degree, betweenness-centrality compactness and closeness-centrality compactness with the effectiveness of attack strategies was analyzed in theory. To validate the proposed assumption, four types of complex networks were selected as test-beds for simulation. The results show that: the effectiveness of attack strategy based on node degree is the worst; with the same average degree, the smaller the betweenness-centrality compactness or closeness-centrality compactness, the more effectiveness of the attack strategy base on betweenness-centrality compactness or closeness-centrality compactness is.