Using ultrashort echo time (UTE) magnetic resonance imaging (MRI), T2∗ can be measured and used to evaluate cartilage degradation. In general, high-quality MRI is required for accurate T2∗ measurement. However, such high-quality MRI requires long scan times which make the clinical utility of T2∗ measurement infeasible. Thus, a method that can produce satisfactory T2∗ maps within reasonable scan times (i.e. limited data) is required. Non-linear mixed-effects model (NLME) is a powerful tool for analyzing incomplete longitudinal data. Since NLME contains both fixed and random effects, it is preferred over more traditional methods. The aim of this study is to investigate whether T2∗ map can be generated using NLME. The simulation results show that NLME is a feasible method for T2∗ measurement. Moreover, NLME can improve the accuracy and precision of parameter estimates compared to the traditional non-linear least square fit.