Abstract: Research on the mechanical properties of chondrocytes under mechanical loading is crucial to understanding their normal and pathological states of chondrocytes and the etiology of osteoarthritis. Based on the highly complex nonlinear relationship between mechanical response and the constitutive parameters of the finite element method computational model of chondrocyte, two inversion methods were proposed to identify the MSnHS constitutive parameters of chondrocytes using the two-way deepnets (TW-Deepnets) model and random forest (RF) model, respectively, combined with the finite element method. Firstly, a three-dimensional finite element model was developed to simulate a stress relaxation unconfined compression test of a single cell. The spatial points of MSnHS constitutive parameters and corresponding finite element compression reaction force-response data were collected. Secondly, the TW-Deepnets model and RF model for chondrocyte parameter inversion were built by combining the Bayes hyperparameter optimization algorithm, and the data collected by the finite element method were trained by the machine learning models. Based on the experimental data of a single chondrocyte subjected to 50% compression, the MSnHS constitutive parameters of chondrocytes were reversed. Finally, the effectiveness of the proposed inversion methods were verified by comparison with the experimental curves from Nguyen, and the accuracy of the two methods was compared and evaluated by determinant coefficient R². The proportion of the importance of each parameter in the MSnHS constitutive model to the mechanical response of chondrocytes was analyzed, and the prediction performance of the two models for each constitutive parameter were tested. The results show that TW-Deepnets model and RF model combined with finite element method are effective and accurate method to identify the MSnHS constitutive material parameters of chondrocytes, and the obtained constitutive parameters can describe the time-dependent mechanical properties of chondrocytes. This method can also be extended to the complex parameters inversion problem of biological cells under static or dynamic loading conditions.