Abiotic features of Antarctic terrestrial habitats, particularly low temperatures and limited availability of liquid water, strongly influence the ecophysiology and life histories of resident biota. However, while temperature regimes of a range of land microhabitats are reasonably well characterized, much less is known of patterns of soil water stress, as current technology does not allow measurement at the required scale. An alternative approach is to use the water status of individual organisms as a proxy for habitat water status and to sample over several years from a population to identify seasonal or long-term patterns. This broad generalization for terrestrial invertebrates was tested on arthropods in the maritime Antarctic. We present analyses of a long-term data set of body water content generated by monthly sampling for 8–11 years of seven species of soil arthropods (four species of Acari, two Collembola and one Diptera) on maritime Antarctic Signy Island, South Orkney Islands. In all species, there was considerable within- and between-sample variability. Despite this, clear seasonal patterns were present in five species, particularly the two collembolans and a prostigmatid mite. Analyses of monthly water content trends across the entire study period identified several statistically significant trends of either increase or decrease in body water content, which we interpret in the context of regional climate change. The data further support the separation of the species into two groups as follows: firstly, the soft-bodied Collembola and Prostigmata, with limited cuticular sclerotization, which are sensitive to changes in soil moisture and are potentially rapid sensors of microhabitat water status, secondly, more heavily sclerotized forms such as Cryptostigmata (=Oribatida) and Mesostigmata mites, which are much less sensitive and responsive to short-term fluctuations in soil water availability. The significance of these findings is discussed and it is concluded that annual cycles of water content were driven by temperature, mediated via radiation and precipitation, and constituted reliable indicators of habitat moisture regimes. However, detailed ecophysiological studies are required on particular species before such information can be used to predict over long timescales.