Abstract: Aiming at the problem of recognizing the parameters of the motion model of unmanned boat,taking the doublepropeller-propelled unmanned boat as the research object,considering the influence of uncertain environmental perturbation on its motion state,and in order to improve the accuracy and anti-perturbation of the recognition algorithm,the paper proposes an interactive recognition method based on the Extended State Observer.Firstly,according to the idea of separation-type modeling,the response characteristics of propeller speed to the command signal are introduced to construct the three-degree-of-freedom dynamics equation of the double-propeller-propelled unmanned boat,which is parameterized to get the linear model used for the identification.Then based on the least squares criterion function,it is proved from the perspective of mathematical expectation that the interaction identification method introducing perturbation estimation can improve the identification accuracy.An expansion state observer is designed to estimate the unknown perturbations in the model,and the estimated value is used as the a priori information of the multi-neo-interest recursive least squares identification algorithm,and the obtained identification value will be used as the model basis for the next perturbation estimation,which is interactively updated on-line and improves the immunity to perturbation of the identification algorithm.Finally,the effectiveness of the interactive discrimination algorithm is verified by real ship test experiments.