Due to the impact of crude oil prices on refinery revenues, the petroleum industry has switched to processing low-cost crude oils. They are blended with high-quality crude oils to feed the crude distillation units (CDUs) with reliable feedstock. The blending process takes place in storage tanks receiving crude parcels from ultralarge carriers and by mixing feed streams supplied to CDUs from multiple tanks. The crude blending and scheduling problem is usually represented by a large nonconvex Mixed-integer nonlinear programming (MINLP) model. This work introduces an effective MINLP continuous-time formulation based on global-precedence sequencing variables to arrange loading and unloading operations in every tank. In addition, synchronized time slots of adjustable length permit to model the sequence of feedstock for each CDU. The basic solution approach consists of sequentially solving a very tight mixed-integer linear programming (MILP) model and a nonlinear programming (NLP) formulation that uses the MILP-solution as the starting point. For large problems, it has been developed a novel solution strategy that incorporates a time-partitioning scheme using the notion of vessel-blocks. The proposed approach has been applied to a series of large examples studied by other authors finding near-optimal schedules at much lower CPU times.