Wildfires have a significant impact on the Earth’s vegetation, atmospheric composition, and climate. There is also growing evidence that fire behaviour has already been altered in response to climate change. Given the anticipated increase in climate conditions conducive to wildfires in many regions of the world, there is an urgent need to advance our understanding of the drivers of wildfires and improve their representation in global Earth system models. Recognising and predicting such responses to climate change and the associated feedbacks is one of the key challenges of the field, and integrated vegetation–fire models have the potential to accomplish this goal. However, the relationship between wildfire activity and past vegetation productivity is still unclear. It has been hypothesised that this strongly contributes to the poor performance of state-of-the-art fire models in representing observed vegetation–fire relationships. Thus, this thesis examines the role of seasonal and long-term vegetation dynamics (and fuel accumulation dynamics) that are of fundamental importance for global wildfires due to their impact on fuel availability during the fire season. In addition to instantaneous climatic conditions, the seasonality of antecedent vegetation and climate conditions controlling fuel build-up and fuel drying was found to be important for the prediction of burnt area in an empirical analysis of wildfire drivers. These were then used to modify the vegetation parametrisation of the INFERNO fire model. By using sigmoidal relationships to explicitly consider antecedent vegetation and climate conditions, global performance was improved. Finally, the new model was evaluated using sensitivity analyses, demonstrating the overarching importance of dryness and temperature while also disentangling the different impacts model parameters have on the magnitude and phase performance of the new parametrisation.