In the present study the deterministic thermo-chemical pultrusion simulation of a composite rod taken from the literature [7] is used as a validation case. The predicted centerline temperature and cure degree profiles of the rod match well with those in the literature [7]. Following the validation case, the probabilistic design of the pultrusion process, which has not been considered until now, is performed. The effect of statistical variations in the material (i.e. fiber and resin) and resin kinetic properties, as well as process parameters such as pulling speed and inlet temperature on the product quality (degree of cure) are examined by means of Monte Carlo Simulation (MCS) with Latin Hypercube Sampling (LHS) technique. The variations in the activation energy as well as the density of the resin are found to have a strong influence on the centerline degree of cure at the exit whereas the other process parameters have smaller influences. Moreover, different MCS options are examined to investigate their effects on the accuracy of the random output parameter.