An approach is presented for the robust stacking sequence design of composite plate wings with uncertain ply orientations. An aeroelastic model is constructed using the Rayleigh-Ritz technique coupled with modified strip theory aerodynamics. Gaussian processes are used as emulators for the aeroelastic instability speed in order to efficiently quantify the effects of uncertainty. The critical instability speed is discontinuous as a result of the different potential instability mechanisms, therefore multiple Gaussian processes are fitted to ensure computational efficiency. An order of two magnitude reduction in model runs is achieved for the majority of examples, and an order of magnitude reduction is achieved when a switch between flutter modes occurs. The emulators are used to estimate the probability that instability occurs at a given design speed, which is minimized using a genetic algorithm. Results are compared to deterministic optima for maximal instability speed. Two lay-up strategies are undertaken, a first in which ply orientations are limited to 0°, ±45° and 90°, and a second in which values of ±30° and ±60° may also be taken. Improvements in reliability of at least 85% are achieved. The inclusion of ±30° and ±60° plies enables a 1.7% increase in the nominal instability speed, and an increase in reliability of at least 59%.