Aiming at the reentry trajectory reconstruction problem of reusable launch vehicle (RLV), a fast solving method based on variable trust region sequential convex programming (SCP) was proposed. First, the non-convex trajectory optimization problem was convexified by discretization and linearizing the non-convex constraints. Then the convex optimization problem was solved using the SCP method. In the initial iteration of SCP, a predictor corrector algorithm was applied to design the initial guessing trajectory and determine the terminal time. In the subsequent iteration, a variable trust region strategy was proposed based on optimization performance indexes, which improved the convergence performance of the algorithm. On the basis of the fast solving method, considering the unexpected events that may occur during the RLV reentry process, such as large deviation and target point changing, the trajectory was reconstructed online taking the changed initial and terminal conditions into account and was tracked effectively using LQR (Linear quadratic regulator) method. Finally, the designed method was compared with the Gauss pseudospectral method and traditional SCP algorithm. Simulation results show that compared with the pseudo-spectrum method and the traditional SCP method, the variable trust region SCP method greatly improved the real-time and convergence of the trajectory solution, and has the ability to be applied to online trajectory reconstruction. Besides, the proposed online trajectory reconstruction method has good robustness and immunity.