Due to increasingly frequent unseasonable temperature fluctuations during the dormant season, the grape and wine industries face significant threats under climate change. Late frost in spring and extreme low temperatures in winter kills young shoots and threaten the survival of the whole vine, respectively, impairing vineyard productivity. A nuanced understanding of grapevine dormant season physiology and the development of novel management methods to enhance viticultural resilience in the dormant season are pivotal for the sustainability of grape and wine industries in a changing world. The objectives of this dissertation are: 1) to elucidate the genetic control of grapevine dormant season physiology; 2) to assess deacclimation across diverse grapevine populations; 3) to develop and evaluate novel management methods to improve viticultural resilience in the dormant season; 4) to create and deploy a robust, AI-empowered grapevine bud cold hardiness prediction model to support vineyard viability assessment in cool climate viticultural regions in North America. Methodologically, this study employs comparative transcriptomics combined with experimentation on dormant buds in controlled conditions to unravel the molecular mechanisms of cold acclimation and dormancy transition. Over three years, deacclimation rates and budbreak timing were measured in five Vitis populations with diverse genetic backgrounds to facilitate the identification of genetic variations controlling deacclimation rate and budbreak timing through quantitative trait loci (QTL) analysis. The impact of exogenous abscisic acid (ABA) on the grapevine bud transcriptome during cold acclimation and deacclimation was evaluated, and the feasibility of using a synthetic ABA analog, tetralone-ABA, as a sprayable product to enhance dormant season resilience was tested. Furthermore, a new grapevine bud cold hardiness model, NYUS.2, was developed using 10,157 data points and 72 temperature-derived features through automated machine learning. This model was deployed in two web applications for global historical analysis and regional real-time monitoring of grapevine bud cold hardiness. By integrating genomics, transcriptomics, grapevine physiological evaluation, development and evaluation of vineyard management tools and AI-empowered modeling, this work paves the way for more resilient viticultural practices to help maintain the sustainability of the grape and wine industries in an era of climate uncertainty.