This study explores the role of implied volatility (IV) in volatility combining forecasts for S&P 500 and DAX markets. A range of GARCH models, ad hoc models and STES models were developed to identify the best performing model that served as a base model for subsequent combining process, of which GJRGARCH model appeared to be the superior model among all the individual models. A total of eight combining models based on Bates and Granger as well as Granger and Ramanathan theories were designed to examine the efficiency of IV in the volatility combining forecasts. The empirical results suggest that IV improves the predictive power of volatility combining forecasts, but the assigned weights were seen less inclined to IV. The inclusion of constant and unconstrained models performed the best. Besides, the study has also addressed the robustness of STES models in volatility forecasts. MCS tests were conducted and further reinforced the superiority of the empirical results.