Seaweed farming in the Gulf of Maine has expanded rapidly over the past decade. As part of ARPA -E’s MARINER program, we conducted a selective breeding program to improve the productivity and composition of sugar kelp which could serve new markets for food, animal feeds, bio-products and eventually biofuels. Our population genetics studies of sugar kelp prompted the development of two breeding programs: one for Southern New England and the other for the Gulf of Maine. We maintain about a thousand unique gametophytes that can be used as parents for generating crosses. We have sequenced the whole genome for 278 parents and tested their crosses. Kelp crosses were planted in “common garden” farm arrays over three seasons (2018 through 2021) in New Hampshire and Connecticut. Trait measurements and analyses of yield, composition and morphology for 734 family plots and 9,666 individual kelp blades will be presented. One highlight is that several plots exceeded 20 kg/m harvest wet weight with the top plot weighing 28 kg/m or 4 kg/m dry weight – about 4 times the commercial average. We used pedigree, genotypic markers , and harvest assessment data to predict offspring performance, and we improved the efficiency of on-farm testing and phenotyping. Ultimately, we are meeting our goal of selecting sugar kelp with 20% increased dry matter yield per unit area per generation. We also completed an annotated reference genome for sugar kelp that enables the identif ication of functional genes and variants, as well as n atural mutations on targeted genes to potentially create non-reproductive sporophytes.