This paper proposes an intelligent reflecting surface (IRS)-assisted wireless inland ship multi-access edge computing (MEC) network architecture with time windows, where UAV is deployed to serve time-constrained unmanned surface vehicles (USVs) with IRS. The task execution efficiency maximization optimization problem is formulated by joint considering IRS phase-shift vector, UAVs hovering coordinates and the task execution indicator. To tackle the formulated challenging problem, a heuristic solution is proposed. First, an enhanced differential evolution algorithm is proposed to optimize UAVs hovering coordinates. Moreover, IRS phase-shift vector and task execution indicator are jointly optimized in an iterative manner by the proposed modified deferred acceptance algorithm. Numerical results verify the effectiveness of the proposed algorithm in comparison with some selected advanced algorithms in terms of task execution efficiency.