Multivariate regression is an estimation of a single model with more than one outcome variable. It uses multivariate probability distribution in both data and variable. It can be both simple or multiple regression. Logistic regression is an estimation of a model with qualitative variable which the value is not continuous.