Drought stress is paramount in limiting worldwide maize (Zea mays L.) production and productivity. In addition, drought has limited expansion of maize production to marginal conditions in areas including western North Dakota (ND) and eastern Montana (MT) where short-season drought tolerance is essential. The development of new cultivars in these dryer areas has been slow in part due to unavailability of reliable field phenotyping methods for breeding purposes and existing insufficient information on the genetic control of agronomic and grain quality attributes under drought stress. Our objectives were (1) to investigate root and stalk quantitative traits as novel field phenotyping methods for drought resistance, (2) to examine the variability and genetic control of agronomic and grain compositional attributes across soil moisture regimes, and (3) to discuss implications of findings for developing new maize cultivars under drought stress environments. Ninety-four partial diallel crosses including 47 diverse maize inbred parents and checks were tested in 12 water stress (WS), well-watered (WW) and random drought (RDT) environments in 2013 and 2014. A new high throughput non-destructive maize phenotyping method for drought tolerance is proposed, including the quantitative traits visible brace root number count and its spread width. Additive gene action was predominant in most of the agronomic and grain quality attributes, while non-additive gene action was important for grain yield, root lodging, stalk diameter, and grain oil content. Root and stalk lodging were negatively associated (P < 0.01) with brace root count and spread width. The estimates of narrow sense heritability were higher for brace root spread width (0.23) and number of brace root count (0.24) as compared to grain yield (0.06) and root lodging (0.1) in WS environments. High relative narrow sense heritability (> 0.40) was estimated for grain compositional attributes except for grain protein and amino acids. Stress environments were largely associated for grain oil and total and extractable starch contents (> 0.60). A reciprocal recurrent selection program can be vital to develop cultivars with high grain yield and quality in drought stress.