The inter-cluster separation (ICS) algorithm adds the separation item into the objective function to minimize the fuzzy Euclidean distance and maximize the inter-cluster separation. However, ICS is sensitive to noisy data, so an improved inter-cluster separation (IICS) algorithm is proposed to deal with this problem. It is claimed that IICS is an incorporation of ICS and improved possibilistic c-means (IPCM) clustering. IICS can produce both possibilities and memberships simultaneously, and it overcomes the noise sensitivity problem of ICS and the coincident clusters problem of possibilistic c-means (PCM) clustering. Further, IICS does not depend on the parameters that exist in IPCM. The experimental results show that IICS compares favorably with ICS.