Vibration effects acting in the science light path reduce the performance of the adaptive optics systems (AO). In order to mitigate the vibration effects and to improve the performance of the AO systems, an adequate model for the vibration in necessary. Traditionally, those vibrations are modelled as oscillators (with or without damping) driven by white noise. In this work, we address the identification of a continuous-time oscillator from discrete-time samples of the position. To this end, we use Maximum Likelihood estimation method to estimate the vibrations frequency.