Due to the considerable socio-economic and ecological impacts of the climate, many efforts have been made to predict year-to-year climate variability (e.g., ENSO) and climate changes over the past decades. In particular, advanced data assimilation techniques have been applied to weather-climate forecasting systems in order to generate realistic ensemble initial conditions for improved predictive skill. However, major challenges remain in reducing uncertainties in the initial conditions and errors in forecast models and external forcing, and in assessing predictability limit associated with multi-scale climate processes and nonlinearity. In this context, Advances in seasonal to decadal prediction have been selected as the topic for this year’s international conference of CAS-TWAS-WMO Forum on Climate Science (CTWF).