The present paper has discussed how to improve our real-time tsunami prediction method by the inverse analysis of tsunami profile observed at offshore points. The inverse analysis of the offshore tsunami data estimates the initial displacements of water surface in the segmented tsunami sources. The estimated initial displacements predict tsunami profiles at inshore points of interests. Our prediction method is tried to improve by the introduction of correlation factors with surrounding segmented sources, the supplement of another observation point, or the employment of wider segmented source area. The paper also confirms that the real-time prediction can be applied to predict the tsunami at Shimotsu Port in Wakayama Prefecture as well as in Osaka Bay.