Near Earth Objects (NEO), commonly known as asteroids, are always moving in outer space. These objects could carry very important knowledge or harmful substances. Knowing the whereabouts of an asteroid makes observation easier since asteroids are moving and, in a limited time, an asteroid will pass the observable distance from the Earth. This study has compared the minimum distance prediction of near-earth objects and the Earth using artificial neural networks, machine learning, and multiple linear regression techniques. The models used in the study are the Multiple Linear Regression Model, Feed Forward Neural Network, and Support Vector Regression Model. The study has used a secondary dataset provided by the “Center for Near Earth Object Studies” (CNEOS) project of NASA. Using every method, two models were trained for each. Every model 1 contained all the variables and every model 2 contained three dependent variables. For model 2, dependent variables were reduced by the assumptions used in linear regression. Even though the linear assumptions were not used on the neural networks or machine learning algorithms, every model 2 showed a significant accuracy increase after variable reduction. Model performances were assessed by multiple prediction error values and R-squared values.