Due to relatively high location accuracy, fingerprint-based localization is one of the best approaches in indoor localization that depends on received signal strength (RSS) from different Wi-Fi based indoor environments. Nowadays, Indoor positioning systems (IPS) are used in a wide range of daily life applications that need direction and navigation services. For example, IPS can play an essential role in the lives of people with vision impairment, who may become lost or disoriented in unfamiliar buildings, or need emergency healthcare services. In this paper, we propose the K-Means-Jensen-Shannon divergence, which is the original k-means algorithm extended into a meta-algorithm. The Bregman divergence is a versatile family of distance measurements that unifies the statistical entropic measures with the quadratic Euclidean distance. Nevertheless, the Bregman divergence is asymmetric; we took the right-sided and the left-sided data to symmetrize the centroid as the minimizer of the average Bregman divergence. To validate our proposed algorithm, the results were compared with the traditional k-mean and the affinity propagation algorithm. Our results indicate that our integrated system outperforms other results with around 1m accuracy in an academic building.