Various factors influence animal movements in fragmented landscapes, and determining these factors is key to understanding ecological processes at a landscape scale. My goals were (1) to determine what factors influence movements of Keel-billed Toucans (Ramphastos sulfuratus) in a fragmented landscape in southern Mexico and (2) to use this information to predict how movement patterns might change if the landscape was altered. I developed a cost-distance geographic information system model that adjusts Euclidean distances by a cost of moving through a certain habitat type, Cost was based on habitat preferences exbibited by toucans, I then used this model to predict how movements might be affected by removal of isolated trees and living fences from the pasture matrix and by removal of forest remnants, Toucans moved more frequently between remnants separated by a low cost-distance value. There was a cost-distance threshold beyond which movements between remnants were rare. Below this threshold, fruit abundance influenced toucan movements but remnant area was not influential in that toucans did not preferentially move to large patches. Remnants close to various other remnants were more frequently visited by toucans, indicating that landscape connectivity Influences toucan movements. Toucans incurred a 10-30% cost increase when moving in computer-simulated landscapes, indicating that changes in forest cover or configuration of habitats may negatively affect toucan populations, assuming that increased cost has a fitness consequence. Cost-distance modeling has been relatively unexplored and may be a valuable tool for determining how the configuration of a landscape impedes or facilitates animal movements.