This paper proposes a mid-course guidance algorithm for dual-pulse rocket air-to-air missiles based on the trajectory optimization framework. Pseudospectral Sequential Convex Programming (PSCP) is utilized to solve the trajectory optimization problem for dual-pulse rocket missiles with control inputs of the angle-of-attack and ignition time of the second pulse. Multi-phase pseudospectral discretization enables us to treat the ignition time control problem by introducing an optimization parameter corresponding to the time interval between two pulses. Convex sub-problems are composed to solve the non-convex discretized trajectory optimization problem by linearizing the nonlinear dynamic constraints with a variable trust-region method and a line search method. Numerical simulations show promising results that the proposed algorithm can provide an accurate optimal solution within a few seconds in MATLAB.