Designing advanced regional turboprop airplanes involves addressing the aerodynamic efficiency, structural mass, fuel burn, longitudinal stability, and field performance within a constrained interplay. The paper investigates three different layouts for regional turboprops: standard wing-mounted engines; rear-mounted engines in which the propulsion system is mounted close to the horizontal tailplane; and three-lifting surfaces with a canard. Variables and design spaces for each layout are characterized via wing planform and horizontal-tail geometry; additionally, canard variables are considered when appropriate. Contrary to the approach of exhaustively searching a design space using a full-factorial design, this research applies the surrogate-based multi-objective optimization approach that employs Latin hypercube sampling, Gaussian-process regression modeling, and NSGA-II algorithm. Mission fuel mass, empty aircraft weight, and Direct Operating Cost proxy are the objectives that need to be minimized under a set of constraints related to static stability, takeoff and landing field lengths, climb time, cruise Mach number, landing-gear clearance in terms of aft center-of-gravity location, and fuel tank capacity. Results show that the rear-mounted propulsion helps achieve greater fuel-efficiency due to better wing aerodynamics, but suffers from additional horizontal-tail penalties. Three-lifting-surface configuration appears to provide optimal balance between fuel burn and performance parameters, provided that the canard placement and relative position of wings are chosen properly. The research concludes that the surrogate-based multi-objective optimization methodology presents a powerful method for preliminary aircraft design, in particular for nonstandard regional turboprop airplanes.