In view of the problem that the existing ship centralized control system can not efficiently handle the rapid growth of electronic equipment and intelligent tasks, the edge computing architecture and task scheduling based on multi-objective optimization were studied. A distributed control architecture of intelligent ship based on edge computing was proposed. Aiming at the scheduling optimization problem caused by many tasks and large amount of computation in the edge computing layer in the architecture, the comprehensive performance of the algorithm in computing task scheduling problem was improved and designed based on non-dominated sorting genetic algorithm Ⅲ(NSGA-Ⅲ) by improving and designing population coding, initialization population strategy, discrete adaptive crossover strategy and discrete adaptive mutation strategy. The simulation experiments show that compared with other algorithms, the average performance of the solution set of the NSGA-Ⅲ-ADO algorithm designed is improved by at least 6.8%, 5.9% and 0.8% in total computing delay, total computing energy consumption and load balancing, and its convergence speed is faster, which can better solve the problem of computing task scheduling.