Genetic relationships among 28 new improved rice varieties were established using Amplified Fragment Length Polymorphism (AFLP) markers. Cultivars were analyzed with 10 EcoR1 and MseI primer combinations. A total of 517 fluorescent AFLP markers were generated and analyzed. Of these 480 fragments were polymorphic (92.84%) and 37 (7.16%) fragments were monomorphic. The Jaccard’s similarity indices (J) based on the AFLP profiles of the 28 varieties were computed and Unweighted Pair Group Method with Arithmetic mean (UPGMA) based dendrogram was constructed. The dendrogram separated varieties into three major clusters. Outliers used in the study were uniquely separated from the rest confirming the reliability of data and analysis. The Cophenetic correlation with 0.862 strongly supported the clustering pattern of UPGMA dendrogram. Principal Coordinate analysis and the unrooted tree also confirmed the clustering pattern of the UPGMA dendrogram. Rice varieties in the same cluster showed similar characteristic features (Eg. Grain colour, life span etc). Therefore this genetic diversity data at molecular level will provide detailed estimates of the genetic variation among Sri Lankan new improved rice varieties and also useful in ex situ and in situ genetic conservation, utilization and exchange of genetic material.