Measurement uncertainty is an important parameter for scientifically characterizing measurement results. This article systematically studies the measurement uncertainty evaluation of linear transfer (LPU), Monte Carlo (MCM), and extended methods of linear transfer. The linear transfer method applies only to linear measurement models. If the measurement model isn't linearized or the output quantity's probability density function greatly deviates from the normal distribution, it may yield unreliable evaluation results. The MCM method utilizes the powerful computing power of computers to obtain the distribution form of the measured data. The covariance matrix method is essentially an extension of the linear transfer method, which can be used to evaluate the uncertainty of linear measurement models with multiple inputs and outputs. Generally, the covariance matrix method preserves the correlation analysis of input variables, which is more in line with measurement reality.