A new service can be constructed1 by composing web services in service-oriented architecture (SOA). Although many selection algorithms have been proposed, a universal one cannot be determined because each algorithm has different characteristics and each situation is unique. Consequently, choosing an algorithm according to the user's requirements leads to selecting the optimum web service. This research develops recommendations of selection algorithms by a quantitative score comparison. We define the quality of algorithm (QoA), which expresses the characteristics of the algorithm, and propose a method to recommend algorithms that satisfy the user's requirements. Our method can compare and select algorithms quantitatively by merely inputting the desired characteristics into the system, reducing the costs of selecting services and improving the user's satisfaction.