This chapter explains several methods that allow readers to estimate the input parameters used by most portfolio optimization models: the expected returns vector and the covariance matrix. These methods mainly use the sample of asset returns as input and, depending on the method, some additional parameters. They are widely used by students, academics, and practitioners because they do not require assuming an explicit relationship with additional variables or the use of prior information like risk factor models and Black Litterman models, respectively. However, their main disadvantage is that the estimated parameters assume that the future performance of asset returns will be like the sample used to estimate them.