This paper presents a two-step framework for the selection of the optimal tourist origins portfolio for a particular destination. The paper applies this decision-making process to French Polynesia. The first step of the framework is based on a mean-variance optimization procedure and proposes the subset of portfolios among which the decision-maker must limit her-his choice. Second, the multi-criteria ELECTRE method is employed to rank all the portfolios considered on the basis of decision-makers' preferences exposed in the parameters of the algorithm. Three decision-maker profiles are proposed from a risk-averse profile to a risk-lover profile. This paper contributes to the dedicated literature by presenting ELECTRE III as an alternative to the utility function approach used by previous studies. Results of our application to French Polynesia's tourist attendance data (from 2014 to 2017) highlight the usefulness of the framework exhibited and empirically underline the economic perspectives offered by Chinese tourists. The French Polynesian application follows a clear presentation permitting consideration of applications to other destinations.