Abstract: In order to execute diversified and complicated on-orbit service tasks, the cellular space robot (CSR) is proposed to perform takeover control for non-cooperative targets. To identify inertial parameters of the non-cooperative spacecraft in takeover control, an interactive parameter identification method for CSRs is proposed to reduce the central node dependency and huge computation requirement of the recursive least squares method which is commonly used. The parameter identification models for mass, center-of-mass and inertia tensor are presented respectively. The method presented is distributed and asynchronous, and requires only a few data interaction. Simulations are conducted to validate the convergence and consensus of the identification results, and high-accuracy identification is realized. More simulations show that the method proposed can achieve convergence and consensus under the influence of measurement noises, communication disconnection, communication topologies and interaction strategies, and obtain accurate results in the case of noises and disconnection. Meanwhile, the simulations validate the superiority of the asynchronous interaction over the synchronous interaction.