Different types of crude oils, including heavy sour crudes, are usually processed in oil refineries. They are suitably mixed in multipipeline crude oil blending systems (MCOBS) to get qualified feedstocks for the crude distillation units (CDUs) that separate them into various fractions or cuts. Current research on the scheduling of such refinery operations seeks to minimize the total operating cost while limiting the concentration of some impurities in the feedstock. However, an essential property like the feedstock composition usually approximated by the true boiling point (TBP) distribution curve is often ignored. In fact, one of the major goals of the blending process is to supply feedstocks to the CDUs that consistently produce the desired proportions of final cuts to match refined product demands. This work introduces a novel slot-based mixed-integer nonlinear programming (MINLP) formulation for the MCOBS scheduling problem that explicitly accounts for the TBP curves of the crude oils to get a series of suitable and stable feedstocks for the CDUs. An MILP-NLP solution strategy is proposed to find near optimal schedules in reasonable CPU times.