This thesis investigates the closed-loop control of continuous-time dynamical systems affected by a random input delay. Stochastic delays are indeed commonly used to represent packet loss and reordering which often take place in network systems, for instance. Yet, stabilization tools for continuous-time processes subject to continuous-time stochastic delays are lacking in general in the literature. This thesis thus proposes a generic stabilization methodology relying on prediction-based controllers, which have strong merits in the deterministic delay case. Modeling the input delay as a Markov process with a finite number of states, we study different prediction-based controllers and, generally speaking, prove that mean-square closed-loop st...