Abstract: Indoor navigation system is the one of interesting application among the researchers in an indoor environment due to the meter-level accuracy requirement in complex structure. This research proposed an improvement of the indoor navigation system based on fingerprinting technique by using K-Means (KM) clustering algorithm. The unknown positions are estimated by using Least Square (LS) and K-Nearest Neighbor (KNN) algorithms. The experimental results show the performance comparison between no-clustering case and KM-clustering case. Finally, we found that the KM clustering algorithm can be improved the accuracy of indoor navigation system both LS and KNN algorithm.