Agroforestry is gaining interest for its potential in carbon sequestration, thereby enhancing climate resilience and contributing to the long-term sustainability of agricultural systems. In Tunisia, the olive tree plays a central role, occupying nearly half of the arable land and over 80% of arboricultural areas. Traditionally, olive groves supported rich biodiversity due to low-intensity agricultural practices and greater diversification within a mosaic system in which the olive tree was the main component. However, this biodiversity has significantly declined in recent decades as a result of modernization, characterized by the expansion of monoculture and the intensification of chemical inputs. These intensified practices not only undermine the health of agroecosystems but also contribute significantly to greenhouse gas emissions, thereby exacerbating climate change. In response to these challenges, this study proposes an ecological intensification model for olive-growing systems, specifically adapted to the arid conditions of central Tunisia. The model emphasizes high species diversification as a strategy to counteract the adverse effects of modern intensive olive farming. The Delphi method, involving a wide range of experts and stakeholders, was used to conduct a SWOT analysis and reach a consensus on strategic directions. The SWOT analysis highlighted agroforestry's main strengths as its ability to adapt to climate change through the creation of micro-ecosystems and its contribution to improving soil fertility. Weaknesses include the potential spread of diseases and water competition arising from the integration of unsuitable crops. Opportunities lie in the introduction of high-value crops and market diversification. The primary threats remain water scarcity and climate-related risks. Recommended associations include annual and forage crops, diversified livestock systems, and arboriculture. These models may serve as a basis for developing effective olive-growing diversification strategies in arid and semi-arid regions.