In this article, conflict probability between aircraft in uncontrolled airspace is estimated. For that purpose, one reviews how to characterize aircraft trajectories with stochastic processes. Indeed, Markov chain model enables us to simulate aircraft trajectories. Conflict probabilities can be estimated with Monte Carlo simulations. Nevertheless, since Monte Carlo methods are not efficient to estimate small probabilities, one proposes the use of the important splitting method. This algorithm is applied on realistic situations of aircraft conflict.