The initial of work presents a discussion about the Monte Carlo Method and about the generation of samples of random variables and ran dom vectors. Generators based on linear congruential, on the Inverse Trans form Method and, on Metropolis-Hastings algorithm (Markov Chain Monte Carlo Method MCMC) are analyzed. Later,techniques for vibrations ana lysis of single-degree of freedom, multi-degree of freedom and continuous stochastic mechanical systems are presented. The system properties are con sidered dertministic and, the forces are characterized by random process. Throughout the work sections, it is show how the system´s displacement response (mean, autocorrrelation and spectral density) can be characterized from the force informations. The Fourier transform, the spectral density function and transfer function are powerful tools used in the study. In the case of multi-degree of freedom systems, the analysis is done by two methods: Impukse Response and Normal modes. For continuous systems, it is made by Monte Carlo method. An example of one continuous mechanical system with two random loads were developed. In this example, MCMC is used to generate samples of the forces. For each realization, the problem is discretized by the Finite Element Method and one approximation of the solution is obtained by the Galerkin Method. In this part of the work, the weak formulation of the motion equation of the vibrating beam and the reduced model are presented. Throughout the work many rountines in MATLAB were developed.