The eigenvector spectrum method is an effective method for resolving coherent signals in array processing. It requires that the minimum eigenvector of the received array signals covariance matrix R should be obtained first. Instead of gradient descent and space triangular ladder methods，we propose a new one-transforming adaptive algorithm method for implementation of the eigenvector spectrum. This method turns the problem of finding the minimum eigenvector，through a simple transformation，into that of finding the maximum one，thus the complicated calculation of R^-l can be eliminated. Theoretic analysis and simulation results show that this new method is superior to gradient descent one in convergence rate and to space triangular ladder one in computation complexity.