This paper considers a WiFi based indoor localization algorithm for heterogeneous networks (HetNets). The given localization algorithm is based on the better modeling of the signal propagation in the wireless environment and proposed in the literature for homogeneous networks. Those studies prove the performance improvement of the lognormal mixture shadowing model over the classical shadowing models for homogeneous networks. Based on this, we extend those results to heterogeneous networks and test the proposed algorithm with experimental setup. The testbed implementation confirms that the taking the information from multiple networks into account shows better localization performance than the individual networks.