This thesis presents new research on the luminescence dating of burnt mounds in the Orkney and Shetland Islands. Through an examination of available evidence for Scottish burnt mounds, a number of key problematic areas have been identified in relation to our understanding of these sites and their place within the archaeological record. Previous chronological investigations of burnt mounds have so far provided little information on the likely duration of individual sites. Site information processes are likewise poorly understood. Luminescence dating has been outlined as a method suited to determining the age of both excavated and unexcavated sites, and to tackling issues of site formation. A combination of stratigraphic and surface sampling at sites on the island of Eday, Orkney and at coastally eroding sites across Shetland has provided suitable material for study. In addition, detailed sampling during excavation of Cruester burnt mound, on Bressay, Shetland has enabled the collection of a series of samples directly linked with the formation of the mound and structures at the site. Fieldwork is reported together with detailed characterisation of the external and internal dose rates of samples collected. Additive dose thermoluminescence dating techniques have been applied to extracted feldspar grains. A procedure for correcting temperature-shift due to thermal contact variation has been developed and implemented, leading to improvements in data processing. Problems have been identified relating to unequal sensitisation at different stages of equal-predose additive dose run which cause normalisation errors leading to incorrect dose estimates. Whilst the underlying physical origins of such changes are not yet firmly understood, a correction method based on modelling of the sensitisation behaviour has been applied. When both sets of corrections are applied, satisfactory plateau responses are obtained, and data from controlled experiments are consistent with external controls.