Despite significant advances in subseasonal-to-decadal prediction, predicting the East Asian surface climate one season ahead remains a formidable challenge. This study proposes a hybrid dynamical–statistical model to improve East Asia’s springtime surface air temperature (SAT) prediction by incorporating the Quasi-Biennial Oscillation (QBO) forecasted by decadal climate prediction systems. The QBO, a dominant interannual variability in the tropical stratosphere, shows a significant negative correlation with the East Asian SAT in March. During the westerly QBO phase, the Asia–Pacific jet shifts equatorward and induces a large-scale cyclonic circulation anomaly over the North Pacific. This leads to a cold SAT anomaly in East Asia, especially in western and central China. The decadal climate prediction systems reliably predict the QBO up to 18 months in advance but have poor prediction skills for the East Asian SAT, even at a few months’ lead time. The hybrid dynamical–statistical model, which combines the QBO forecast from decadal climate prediction models and the QBO–East Asian SAT relationship in the observations, significantly improves the East Asian SAT prediction at a 5-month lead time. This approach suggests that improving the multi-year QBO prediction may enable the East Asian SAT prediction even one year in advance.