The fragmentation of agricultural land is influenced by political, economic, social, ecological and environmental factors, which affect its dynamics, patterns, structures, and functions. However, a deep analysis aimed at examining agricultural land fragmentation and its driving forces, and predicting future agricultural land fragmentation is needed. We investigated the degree of fragmentation in a case study in Torres Vedras municipality (Portugal). This territory has experienced significant agricultural land use changes in the last twenty years, mainly in the form of agricultural intensification and land fragmentation. The purposes of the study are: (1) to identify different degrees of agricultural land fragmentation; (2) to analyse and recognize underlying driving forces; (3) to identify the effect of scale; (4) and to predict agricultural land fragmentation for 2025 in a business as usual scenario. This approach concentrates on a cluster analysis to define the agricultural land fragmentation and a multi-layer perceptron to project agricultural land fragmentation. The results indicate that a 5 by 5 km scale of analysis is more efficient than 2 by 2 km to identify the most influential driving forces, in which human activities are one of the main causes of agricultural land fragmentation. In addition, the results also predict that agricultural land fragmentation will increase in 2025. These outcomes should be used to forecast agricultural land fragmentation and to reduce negative impacts.