Hierarchical identification is an important branch of system identification.The hierarchical identification principle is developed on the bisis of the "decomposition-coordination principle" in the hierarchical control for a large-scale system.It is able to not only solve problems that the identification algorithms require heavy computational burden for a lareg-scale systems with many parameters and high dimensions problem,but also solve identification problems for bilinear-parameter systems,multi-linear-parameter systems and nonlinear systems with complex structures.In this paper,firstly we describe the hierarchical identification principle,the Jacobi iteration and Gauss-Seidel iteration for linear systems with a set of equations Ax=b,and give the family of iterative methods for linear equations;secondly,we study hierarchical least squares based and hierarchical gradient based iterative algorithms for general matrix equations and coupled matrix equations in the light of the Jacobi iteration and the hierarchical identification principle;thirdly,we present a two-stage recursive least squares algorithm(i.e.,a simple hierarchical least squares algorithm) for equation error models and a hierarchical least squares identification algorithm for linear regression models.Finally,the hierarchical identification methods are introdiced for multivariable CARMA-like systems using the hierarchical identification principle.