特征空间方法是现代谱估计的重要方法之一。针对噪声中的复正弦信号的模型，通过特征矢量的方法进行信号空间的分解。通过将信号空间分解为谐波信号空间和噪声信号空间这两个相互正交的空间的方法进行谱估计。Pisarenko,Music,ESPRIT算法都是这一类方法的重要算法。我们通过对不同算法进行比较分析不同算法在各种状况下的表现性能。并针对ESPRIT算法的各种改性和提升进行对比和测试。-Feature space is one of the modern spectral estimation method. Complex sinusoidal signal for the noise model, by the method of the feature vector for the decomposition of the signal space. By decomposition of the signal space to the space harmonic signal and the noise signal space, the method of the two mutually orthogonal spatial spectrum estimation. Pisarenko, MUSIC, ESPRIT algorithm is an important algorithm of this type of method. Different algorithms, our comparative analysis of the performance of different algorithms performance under a variety of conditions. And for a variety of modification and upgrading of the ESPRIT algorithm to compare and test.