In research and statistics projects, if we know the form of the regression, parametric inferences often perform better than semiparametric inferences and nonparametric inferences. This talk focus on the maximum likelihood estimator (MLE) under the uniform distribution with linear regression data. We first introduce and discuss the assumptions and the conclusions in Robbins, H. and Zhang, C.-H.'s paper. Then we develop the closed form algorithm to find the MLE under the multiple linear regression model. Some results on this subject with the simple linear regression and non-random covariate are introduced and compared.