The Analytic Hierarchy Process (AHP) with multiple experts usually aggregates the pairwise evaluations with geometric means.Depending on the data set asymmetry, aggregation by average can cause distortions.In addition, there is an uncertainty associated with human judgment, due to cognitive limitations, variations in knowledge and experience, and bias, which imprecise the process.To deal with these difficulties, Monte Carlo simulation has been associated with AHP, creating MCAHP models with stochastic results.This form of result fits the first step of the Composition of Probabilistic Preferences (CPP), a probabilistic multicriteria decision aid method.The association of CPP improves the interpretation of results and extend the MCAHP to different points of view of decision making.A new model is developed here, the CPP-MCAHP, which is compared to AHP with geometric means and stochastic AHP with satisfactory results.