MVSCCPO is a mean-variance-skewness portfolio optimization problem with cardinality constraint. It increases the skewness as the third objective, and introduces cardinality constraint to reduce the difficulty of asset management. A customized NSGA-III(c-NSGA-III) algorithm is proposed for the problem. It introduces an effective solution representation mechanism, customized crossover and mutation operators, a tournament selection based on membership is introduced in the mating selection phase instead of the original random selection method. Finally, an empirical analysis is carried out to verify the effectiveness of the proposed algorithm by comparing with NSGA-II and MOEA/D. The results show that c-NSGA-III performs better than other algorithms on solving MVSCCPO problem.