Dengue fever (DF) is one of the most common mosquito-borne viral diseases in the world. Typical symptoms of (DF) include fever accompanied by headache, muscle pains, rash, cough, and vomiting. A proportion of cases can progress to severe dengue, known as dengue haemorrhagic fever (DHF), associated with increased vascular permeability, thrombocytopenia, and haemorrhages. To date, there are no accurate means to predict which patients will progress to severe dengue at first presentation, meaning that in endemic areas, health facilities are overwhelmed with patients admitted for observation, posing a huge socio-economic burden on health systems. Strategies aimed at predicting disease progression are urgently needed to improve patient management. To address this issue, we pursued a systems immunology approach integrating plasma cytokine profiling, high-dimensional mass cytometry and peripheral blood mononuclear cell (PBMC) transcriptomic analysis in a prospective study of individuals from a dengue-endemic area of Indonesia, that progressed to develop either uncomplicated DF or DHF. Elevated levels of inflammatory chemokines as well as high frequencies of non-classical monocytes were associated with progression to DHF. Furthermore, circulating levels of CD4+ non-classical monocytes predicted risk of severe disease at the onset of fever, after a primary and secondary dengue virus infection. Progression to severe dengue was also associated with an important transcriptional signature featuring impaired T cell activation and cell division as well as reduced glycolysis and oxidative stress response. In contrast, protection from DHF was associated with high frequencies of CD4+ and CD8+ effector memory T cells expressing elevated levels of the co-stimulatory molecules ICOS and CD27, as well as high numbers of CXCR3+ TH1-polarised T follicular helper cells. Our research demonstrated that efficient effector T cell memory activation plays a critical role in protection from DHF during secondary dengue infections. The results also provide proof of concept for the potential of system immunology approaches to identify discrete populations in the blood predicting increased risk of DHF to develop diagnostic tools for early detection of complicated cases at point of care.