针对软件白盒子安全问题，提出了一种基于参数化分解树的控制流二次平展混淆的改进方法． 该方法首先根据设定的深度、广度及粒度的上界构建参数化分解树，然后用一个 while-switch 循环选择结构统筹整棵树，并对树中满足一定条件的节点进行二次平展． 实验结果表明:与基于参数化分解树的控制流平展混淆方法相比，文中方法减少了执行开销和解决深层不作为问题;与传统的控制流平展混淆方法相比，文中方法增加了反编译及逆向工程的难度．

Aiming at the security problem of software white box，an improved parameterized decomposition tree-based obfuscation method with double flattening control flow is put forward． On the basis of given upper bounds of depth，breadth and granularity，the method builds a decomposition tree，coordinates the whole tree with a cycle se-lection structure named while-switch，and then applies double flattering to relevant nodes that satisfy certain condi-tions． Experimental results indicate that，in comparison with the flattening obfuscation method of control flow on the basis of parameterized decomposition tree，the proposed method reduces the execution expense and solves the deep nonfeasance problem; and that，in comparison with the traditional method only with flattening control flow，the pro-posed method increases the difficulty in decompilation and reverse engineering．