The engineers and scientists want mathematical models of the observed system for understanding, design and control. Most of these systems are nonlinear. There is not a unique solution because of the many different types of nonlinear systems with different behaviors and so the modeling is very involved and universally usable design tools are not available. For these reasons the nonlinear systems are often approximated with linear systems, because this theory is user friendly and well understood. In this paper we will discuss the best linear approximation in least squares sense.