This paper proposes to identify Agricultural futures' roles (Safe-haven, Hedge, and Diversifier) in the Thai stock market during 2000-2020 by applying a bivariate Cross-Quantilogram (CQ) approach. The CQ approach can examine the cross-quantile correlation between assets, while the traditional approaches (such as GARCH, DCC, and MSV) examine only mean-to-mean dependency structures. The CQ methodology can estimate the tail dependencies and directional predictability between financial assets more accurately than traditional methods since financial assets typically have a skewed and asymmetric distribution. The correlation between assets during the extreme market condition (tail dependencies) is important to classify a financial asset as a Safe-haven role. The agricultural commodities considered in this study are the most active asset categories in the markets (cereals, oilseeds, other soft commodities, and miscellaneous commodities). The results show that the agricultural assets are more explicitly correlated with the Thai stock market in crisis periods, such as a negative result in canola during COVID-19. Agricultural commodities, including wheat, oats, and canola, can play a strong safe-haven role in the Thai stock market, according to the lowest cross-quantiles (bearish market) data. According to the results of overall quantiles (normal situations), wheat, corn, canola, soybean, and sugar can all be used as hedges. The rolling windows for directional predictability, which show the time-varying CQ, confirm that these agricultural commodities can be served as Safe-havens throughout the study periods. Therefore, including these specific agricultural commodities (Safe-haven or Hedge) in a portfolio of Thai stocks will help lower risk and boost performance under normal and extreme downturn situations.