Most of the existing multi-objective evolutionary algorithms (MOEAs) are designed to solve many-objective optimization problems (MaOPs) with regular fronts. However, their effectiveness for MaOPs with irregular fronts are yet to be improved. In this paper, a cone decomposition evolutionary algorithm with dominance-based archive (CDEA-DA) is presented to extend decomposition-based MOEAs for MaOPs with irregular fronts. In CDEA-DA, an improved cone decomposition strategy is adopted to decompose one MaOP into several scalar subproblems. Then, a dominance-based archive is designed to collect the non-dominated solutions eliminated during evolution, so as to improve the quality of the obtained front. The proposed algorithm is compared with four state-of-the-art algorithms on unconstrained benchmark MaOPs. Empirical results demonstrate that CDEA-DA achieves the superior quality of fronts.