Current maritime wireless networks face challenges in covering remote sea areas due to geographical limitations. To address the intolerable latency often encountered in long-range maritime communications, this paper proposes a collaborative transmission framework. Unlike traditional wired connections or fixed base stations, we employ an uncrewed surface vehicle (USV) as a mobile fronthaul for cache-enabled buoys deployed along shipping routes. Multiuser multiple-input multiple-output (MU-MIMO) technology is implemented to enhance diversity gain and reduce interference. In this setup, we leverage the tendency of fleet-based marine users to request similar contents, formulating a problem that jointly optimizes the USV trajectory and multicast beamforming to maximize the minimum rate among fleet members. Due to the coupling between trajectory and beamforming in the MIMO rate expression involving matrix inversion and logdet(& sdot;) operations, we propose a trajectory-dependent equivalent rate transformation method, which eliminates the need for a block coordinate descent process. Using the Lagrange dual method, we derive a closed-form solution for buoy network-wide beamforming at each step. Simulations confirm that the USV-enabled wireless fronthaul and collaborative framework significantly reduce transmission latency.