Clustering is a method for parti- tioning a data set into groups so that similar elements are assigned to the same group and di􏰢erent elements are assigned to di􏰢erent groups. In addition to the traditional clustering methods, clustering for probability density functions (CDF) has been widely studied to capture the data uncertainty. In CDF, au- tomatic clustering is a clever technique that can automatically determine the number of clusters. In the existing automatic clustering algorithms, the new probability density function fi (t) is updated based on the weight mean of all previous probability density functions fj(t−1),j = 1,2,...,N, which can lead to slow convergence. This paper proposes an e􏰥cient automatic clustering algorithm for probability density functions. In the proposed approach, the updating of fi(t) is based on the weight mean of {f1(t),f2(t),...,fi−1(t),fi(t−1),fi+1(t−1),... ,fN(t−1)}, where N is the number of func- tions and i = 1,2,...,N. This technique can inherit the new probability density functions that have just been updated, thereby leading to fast convergence. The numerical examples demonstrate the superiority of the proposed approach over the existing automatic clustering algorithms.