Clustering is the main and essentially used method for the automatic information taking out from huge amounts of data. Its task is recognize groups, its called clusters, of indistinguishable objects in a data set. Clustering methods is used in broad area, including database shopping, web inspect, information acceptance, bio technology, and broad others. whenever, if clustering methods is used on real data , a problem that often produce up is that absent values shown in the data sets. For established clustering methods were developed to inspect complete data, there is a need for data clustering method pickup incomplete data. Proceed towards recommend in the literature for modify the clustering algorithms to incomplete data effort better on data set with similer scattered clusters. In this thesis we are description a new proceed towards for suitable for a new use the fuzzy c-means clustering algorithm to incomplete data that takesget hold the scatters of clusters into statement. In this experiment on made by human being and real data sets with various scattered clusters .we show that our proceed out accomplish the another clustering techniq for defective data. ,We explain various cluster reliability functions and modify them to defective data according to the “present-case” proceed it. We inspect the original and the modify cluster reliability functions using the separate results of various artificial and real data produce by various fuzzy clustering algorithms for imperfect data. Therefore both the clustering algorithms and the cluster reliability functions are modify to inperfect data, we should target are finding the factor that is searching for determining the optimal number of clusters on defected data: the modify of the clustering algorithms, the modify of the cluster reliable functions, or the miss of information in the data self. In this research , we present that our capacity are capable to right check the clusters nearly situated to every other and cluster within broad density substance .