Sugar mills in Brazil represent significant capital investments. To maintain appropriate returns on their investment, sugar companies seek to run the mills at capacity over the entire nine months of the sugarcane harvest season. Because the sugar content of cane degrades considerably once it is cut, maintaining inventories of cut cane is undesirable. Instead, mills want to coordinate the arrival of cut cane with production. In this paper, we present a model of the sugarcane harvest logistics problem in Brazil. We introduce a series of valid inequalities for the model, introduce heuristics for finding an initial feasible solution, and for lifting the lower bound. Computational results demonstrate the effectiveness of the inequalities and heuristics. In addition, we explore the value of allowing trucks to serve multiple rather than single locations and demonstrate the value of allowing the harvest speed to vary.