This chapter focuses on the extreme load analysis based on physical bending moment measurements. Of particular interest in a wind turbine system are the extreme events under which loads exceed a threshold, called a nominal design load or extreme load. The chapter presents numerical analysis of extreme loads recorded in the Turbine Bending Moment Dataset and discusses the difference between the spline-based approach and the binning-based approach. The increasing size and weight of components of utility-scale wind turbines escalate the loads and the stresses imposed on the structure. G. Lee et al. present a Bayesian spline method for estimating the extreme load on wind turbines. The Bayesian multivariate adaptive regression splines (MARS) model is basically a MARS model but includes the number and locations of knots as part of its model parameters and determines these from observed data. The Bayesian MARS model treats the number and locations of the knots as random quantities.