In this paper, a new type of fixed-wing vertical take-off and landing, unmanned aerial vehicle (UAV) has been designed. Thrust-vector direct force control has been introduced in three axes to make UAV exhibit superior manoeuverability in transition flight. Considering the characteristics of UAV's dynamic model, which are non-linear, non-affine, and have redundant input, a two-stage progressive optimal control allocation method is developed, which can optimise position and attitude control in synthetical, and motivate effectors to generates desired force and moments. A task-oriented weight selection scheme is proposed to make objective function suitable for different tasks and flight conditions. In addition, a general constraint strategy is designed to guarantee the feasibility of optimal allocation results, which can largely reduce the onboard computation time. Simulations show that UAV can adjust flight attitude and use control effectors in an optimal way, and demonstrating satisfactory tracking of low-speed high-manoeuver flight paths.