The main obstacle in task planning to the manufacturing process in a flexible manufacturing cell (FMC) is the interaction between its components and the amount of produced pieces in it, although it is lower than those produced in dedicated manufacturing cell, there is always a problem in the scheduling of the detailed task planning for each piece. This paper presents a task planning methodology for a FMC using heuristic search with partial results, which is a modification of the Best-First search. This method basis is: to produce n pieces in the best way is necessary to produce n-1 pieces in the best way. This idea reduces the compute time by the elimination of the unnecessary states to produce n pieces, starting from the partial result of producing n-1 pieces. This is the initial state for the n pieces manufacturing.