This paper presents a two-stage meta-heuristic approach to producing weight-optimised solutions needed prior to the detailed finite element analysis of composite wing. Composite wing covers are assumed to take the form of a group of stiffened sub-panels with varying skin and stiffener geometries according to the wing layout and loads. A population of limited solutions satisfying various design constraints was created using layout (skin and stiffener geometry), selected lay-ups, rule based stacking sequence and various assumed loads. The closed form analytical solutions of flat stiffened orthotropic plates are used for calculating buckling reserve factors and strength margins. For each sub-panel, a meta-heuristic rule was imposed to search for a suitable combination of skin and stiffener geometry. The criterion used was minimum weight satisfying laminate continuity accounting for manufacturability. Later, the optimised solutions for each sub-panel are converted into a format supported by the conventional finite element tool (NASTRAN). The use of meta-heuristic approach and their automation in Visual Basic for Applications resulted in fast convergence and potential time-saving compared to genetic algorithms.