Modern agriculture and conventional breeding and the liberal use of high inputs has resulted in the loss of genetic diversity and the stagnation of yields in cereals in less favourable areas. Increasingly landraces are being replaced by modern cultivars which are less resilient to pests, diseases and abiotic stresses and thereby losing a valuable source of germplasm for meeting the future needs of sustainable agriculture in the context of climate change. Where landraces persist there is concern that their potential is not fully realised. Much effort has gone into collecting, organising, studying and analysing landraces recently and we review the current status and potential for their improved deployment and exploitation, and incorporation of their positive qualities into new cultivars or populations for more sustainable agricultural production. In particular their potential as sources of novel disease and abiotic stress resistance genes or combination of genes if deployed appropriately, of phytonutrients accompanied with optimal micronutrient concentrations which can help alleviate aging-related and chronic diseases, and of nutrient use efficiency traits. We discuss the place of landraces in the origin of modern cereal crops and breeding of elite cereal cultivars, the importance of on-farm and ex situ diversity conservation; how modern genotyping approaches can help both conservation and exploitation; the importance of different phenotyping approaches; and whether legal issues associated with landrace marketing and utilisation need addressing. In this review of the current status and prospects for landraces of cereals in the context of sustainable agriculture, the major points are the following: (1) Landraces have very rich and complex ancestry representing variation in response to many diverse stresses and are vast resources for the development of future crops deriving many sustainable traits from their heritage. (2) There are many germplasm collections of landraces of the major cereals worldwide exhibiting much variation in valuable morphological, agronomic and biochemical traits. The germplasm has been characterised to variable degrees and in many different ways including molecular markers which can assist selection. (3) Much of this germplasm is being maintained both in long-term storage and on farm where it continues to evolve, both of which have their merits and problems. There is much concern about loss of variation, identification, description and accessibility of accessions despite international strategies for addressing these issues. (4) Developments in genotyping technologies are making the variation available in landraces ever more accessible. However, high quality, extensive and detailed, relevant and appropriate phenotyping needs to be associated with the genotyping to enable it to be exploited successfully. We also need to understand the complexity of the genetics of these desirable traits in order to develop new germplasm. (5) Nutrient use efficiency is a very important criterion for sustainability. Landrace material offers a potential source for crop improvement although these traits are highly interactive with their environment, particularly developmental stage, soil conditions and other organisms affecting roots and their environment. (6) Landraces are also a potential source of traits for improved nutrition of cereal crops, particularly antioxidants, phenolics in general, carotenoids and tocol in particular. They also have the potential to improve mineral content, particularly iron and zinc, if these traits can be uccessfully transferred to improved varieties. (7) Landraces have been shown to be valuable sources of resistance to pathogens and there is more to be gained from such sources. There is also potential, largely unrealised, for disease tolerance and resistance or tolerance of pest and various abiotic stresses too including to toxic environments. (8) Single gene traits are generally easily transferred from landrace germplasm to modern cultivars, but most of the desirable traits characteristic of landraces are complex and difficult to express in different genetic backgrounds. Maintaining these characteristics in heterogeneous landraces is also problematic. Breeding, selection and deployment methods appropriate to these objectives should be used rather than those used for high input intensive agriculture plant breeding. (9) Participatory plant breeding and variety selection has proven more successful than the approach used in high input breeding programmes for landrace improvement in stress-prone environments where sustainable approaches are a high priority. Despite being more complex to carry out, it not only delivers improved germplasm, but also aids uptake and communication between farmers, researchers and advisors for the benefit of all. (10) Previous seed trade legislation was designed primarily to protect trade and return royalty income to modern plant breeders with expensive programmes to fund. As the desirability of using landraces becomes more apparent to achieve greater sustainability, legislation changes are being made to facilitate this trade too. However, more changes are needed to promote the exploitation of diversity in landraces and encourage their use.