Habitat fragmentation displays a crucial role in conservation biology. Despite this, little is known about the detailed ecological consequences of habitat fragmentation due to the scarce number of controlled experimental surveys. The species–area relationship, a fundamental concept in ecology, requires the understanding of the fragmentation effects in a long term perspective, which turns this task even harder. Here we address the spatial patterns of species distribution in fragmented landscapes, assuming a neutral community model. We study the species area relationship and how its shape changes as the landscape becomes more fragmented. Recent investigations, based on extensive computer simulation, have contributed to establish some definite conclusions in the study of non-fragmented landscapes: the existence of a three-regime or two-regime scenario for the species–area relationship, the emergence of a power-law regime at intermediate scales and the augment of the species–area exponent z with the speciation rate. Despite the recent efforts, some other questions remain, such as the dependence of z in the whole range of the speciation rate. Questions like these are currently debated but generalizations cannot be drawn. This is the first paper, to our knowledge, that uses the coalescence method and neutral theory to examine biodiversity on more complex spatial structures. Our simulation results corroborate that the fragmentation plays a crucial role in shaping the species–area relationship, by determining the existence and extension of the power-law regime associated with small and intermediate areas. On the other hand, when individuals are allowed to disperse over longer distances the species–area relationship now displays the classic triphasic pattern, and the intermediate regime, which is well described by a power-law, is established even for highly fragmented landscapes.