Optimizing the spatial configuration of the marine industry is pivotal for driving the green transition of coastal cities. Using high-resolution geospatial data on micro-level marine-related enterprises， this study employs kernel density estimation to precisely measure urban marine industrial spatial agglomeration at the city level， constructs a green economic efficiency indicator that integrates both land and sea dimensions， and empirically examines the nonlinear relationship between agglomeration and green efficiency as well as its underlying mechanisms. The findings reveal a statistically significant inverted U-shaped relationship between marine industrial agglomeration and green economic efficiency， with pronounced sectoral heterogeneity. The inflection point occurs earlier for the secondary industry， while the tertiary industry has a higher sustainable threshold， and the effect of the primary industry is not significant. Mechanism analyses indicate that scale effects， blue technology effects， and spatial allocation effects are important transmission channels， but their patterns differ markedly across industries. Further moderation effect tests show that government attention to marine economic policy steepens the inverted U-shaped curve by intensifying the marginal diminishing effect in the later stage of agglomeration， whereas digital economic development flattens the curve by alleviating congestion costs. Based on these findings， this study proposes micro-level empirical evidence to overcome the limitations of land-based agglomeration theories and offer decision-making references for formulating differentiated marine spatial governance policies.