The species of spare parts in material system of Iron and steel enterprise are very complex, also there are many influent factors on demand, so it’s very difficult to ensure the requirement planning of spare parts. According to the problem of spare parts requirement planning in iron and steel enterprise’s internal, using fuzzy ISODATA algorithm to fuzzy clustering in the influent factors of spare parts purchasing, then, ensuring the membership degree in every kind of influent factor of every spare parts, and using simulation results to verify the feasibility of algorithm, thus, it can accurately determine the most important influent factors’ kind in order to provide evidence for building the model of requirement planning of spare parts.