We applied W phase inversion for the large (M>7) Indonesian earthquakes that occurred in the period of 2004-2009 by using global data obtained from IRIS DMC. The W phase inversion was made for the PDE location, optimum centroid location and the GCMT location. These three locations gave similar inversion results; both moment magnitude and focal mechanism were generally in very good agreement with the GCMT solutions. Comparison of moment magnitudes Mw indicates that seismic moments from the GCMT and W phase inversions are larger than that from body waves for three events that caused damaging tsunamis (the 2004 Aceh, 2005 Nias and 2006 Java earthquakes). Therefore it is necessary to perform analysis of long-period seismic waves to accurately evaluate the tsunami potential. Tsunami simulation was performed for the 2006 Java, 2007 Bengkulu (mainshock), 2008 Sulawesi and 2009 Papua (mainshock) events for the three different centroid locations and sea floor deformations based on the W phase inversion results. The simulation shows that these three models yields similar tsunami arrival times and heights on the coasts where the tsunami arrival times are more than 30 minutes. Based on the W phase inversion and tsunami simulation results we suggest that the PDE location yields good enough solution so that it can be used in real time situation for the warning purpose. For coastal regions near the earthquake source with tsunami arrives less than 30 minutes, W phase analysis cannot be used for the first tsunami warning. Because W phase analysis provides long-period information on the earthquake source and can be performed faster than the GCMT inversion, it can be used to update the tsunami warning at BMKG (Agency for Meteorology Climatology and Geophysics).