Abstract: An approach to classify human chromosomes through an organizing radial basis function network (RBFN) is proposed. The three characteristics of chromosome, the length of the chromosome, the centromeric index and the density profile, are imployed to be the inputs of the organizing RBFN. With an hybrid learning algorithm, the organizing RBFN is tyained to classify the chromosomes. The relevant study results are of reference value in automatic classification of chromosome.