A fuzzy-based optimal delivery scheduling approach is introduced in this research. In this approach, optimal delivery scheduling is conducted at three different levels considering (1) one driver and one load, (2) one driver and multiple loads, and, (3) multiple drivers and multiple loads. Fuzzy-based clustering methods are employed to classify delivery tasks into driver groups, load groups, and location groups. The optimal sequence and timing parameters of the delivery tasks are identified using the fuzzy-based clustering results and state-space search. The intelligent optimal delivery scheduling system was implemented using Smalltalk.