Leaves are a functionally important and visually conspicuous aspect of plant form. In nature, they present with a great diversity of shapes ranging, for example, from simple poplar leaves to prominently lobed maples through to highly compound tomato leaves. In this thesis, I examine the basis of this diversity using computer simulations. To elucidate the biological determinants of leaf form I rst present three case studies focused on diㄦent aspects of leaf development. The rst simulates leaf and midvein initiation in Brachypodium distachyon, and reproduces detailed biological observations of these processes. The remaining focus on leaf margin development, which is thought to play a primary morphogenetic role in the acquisition of leaf form. Thus for the second and third case studies, I propose a model of leaf margin development for simple leaves in Arabidopsis thaliana and compound leaves in Cardamine hirsuta. To simulate leaf margin development the boundary propagation method is proposed, which simulates the leaf margin and its propagation during development. The models developed using this method are qualitatively consistent with biological observations, and elucidate the role of the margin during simple and compound leaf development. To investigate natural leaf form diversity I propose a geometric model of leaf development based on the three molecularly detailed case-studies. This framework simulates leaf development as an interplay between patterning of the leaf margin, the establishment and growth of veins and the progression of maturation in the leaf blade. Although couched in geometric terms, the method is derived from the molecular level models developed in the three case studies, and is thus biologically motivated. This facilitates a biologically meaningful exploration of the diversity of leaf forms seen in nature using the framework. Additionally, it provides a procedural technique for generating leaf forms for computer graphics purposes.