The application of high-throughput plant phenotyping (HTPP) to continuously study plant populations under relevant growing conditions creates the possibility to more efficiently dissect the genetic basis of dynamic adaptive traits. Towards this end, we employed a field-based HTPP system that deployed sets of sensors to simultaneously measure canopy temperature, reflectance, and height on a cotton (Gossypium hirsutum L.) recombinant inbred line mapping population. The evaluation trials were conducted under well-watered and water-limited conditions in a replicated field experiment at a hot, arid location in central Arizona, with trait measurements taken at different times on multiple days across three years. Canopy temperature, normalized difference vegetation index (NDVI), height, and leaf area index (LAI) displayed moderate to high broad-sense heritabilities as well as varied interactions among genotypes with water regime and time of day. Distinct temporal patterns of quantitative trait loci (QTL) expression were mostly observed for the more dynamic HTPP canopy traits, canopy temperature and NDVI, and varied across plant developmental stages. In addition, the strength of correlation between HTPP canopy and agronomic traits such as lint yield displayed a time-dependent relationship. We also found that the position of some QTL controlling HTPP canopy traits were shared with agronomic and physiological traits. This work demonstrates the novel use of a field-based, HTPP system to study the genetic basis of stress-adaptive traits in cotton, and these results have the potential to facilitate the development of stress-resilient cotton cultivars.