In recent years, the EU Commission has enacted various firefighting policies to combat and diminish the adverse effects of wildfires. The Mediterranean area has experienced an observable extension of its wildfire season, coupled with rapid shifts in fire-weather dynamics, resulting in exceptionally severe wildfire occurrences. As of 2022, the EU has recorded an approximate total burned area of 792,902 hectares, with forests accounting for 66% of this figure (Rodrigues et al., 2023). The main objective of this study is to anticipate extreme wildfire conditions by providing a synthetic product depicting the chances of a fire event starting and escaping containment. To do so, we combined empirical models of ignition likelihood and effectiveness of the initial attack stage. We employed machine learning techniques to calibrate binary regression models using historical wildfire ignition data and geospatial layer depicting the main drivers of ignition and containment, namely accessibility, human pressure on wildlands, fuel moisture and availability. We illustrate our approach along the Mediterranean coastal region of Spain. Our approach enables us to predict wildfire contention capacity under diverse population growth and climate warming scenarios. This strategy aims to improve disaster risk reduction by pointing wildfire management zones and prioritizing intervention in high-risk areas. Results indicate a high predictive ability to model human-caused wildfire ignition (AUC>0.80) but a modest capability to capture the containment capability (AUC≈0.70). Accessibility by road largely controls the spatial pattern of ignition and containment, with dead fuel moisture content modulating the temporal pattern of probability. We further illustrate the approach by providing insights into future SSP (Shared Socieconomic Pathways) scenarios by synthesizing both products into comprehensive management zones (Rodrigues et al., 2022).