Most rosaceous tree fruit have long juvenility and large plant sizes, which makes traditional seedling selection (TSS), relying on phenotypic evaluation alone, relatively time-consuming and expensive. Limited predictiveness of phenotypic information also restricts the accuracy of TSS for traits with low heritability. Marker-assisted seedling selection (MASS) uses DNA markers to provide an early DNA-based evaluation of genetic performance potential of seedlings, with the aim of improving cost and/or genetic efficiency of seedling selection. MASS is still not widely adopted in rosaceous tree fruit breeding despite some successful examples. This review assesses reported MASS successes and identifies key elements and remaining challenges. Suggested solutions to widespread MASS adoption in Rosaceae tree fruit breeding are to (1) provide more breeding-program-specific DNA tests for high-impact attributes, (2) develop approaches to readily identify efficient MASS schemes, (3) increase access to service providers specialized in DNA testing for rosaceous tree fruit breeding programs, (4) obtain funds to initially implement MASS, and (5) develop software tools and provide training to apply DNA information. Overcoming current challenges of implementing MASS is likely to facilitate its adoption in scenarios already proven to be effective: where DNA testing is conducted at an early seedling stage for single or multiple traits without significant interactions between them and where trait loci targeted by DNA tests have a major influence on trait levels.