Due to intense interference from climate change and human activities in recent years, the traditional consistent design methods for calculation of design lowest navigable water level no longer meet the demands. In this study, a hydrological variation diagnosis system has been applied to analysis of annual mean water level series for four major rivers in Guangdong: East River, West River, North River, and Hanjiang River. Results show that moderate or strong variations in river level have been detected at the outlet control stations of these rivers. Then, we applied a synthetic duration curve method to the series of river level under the changing environments, using two separate daily water level sub-series segmented at the change point, and compared the calculations of design lowest navigable water level based on the pro- and post-changepoint sub-series and the whole series. The results indicate that climate change has caused an increase of more than 30% in navigation depth at the Chaoan station while a decrease of more than 30% at other stations. A frequency guarantee rate method was also adopted to diagnose the annual series of characteristic navigation levels, and this revealed change points of different forms and degrees in these series corresponding to different guarantee rates. Thus, our results show the existing problems with the two traditional methods in calculation of inconsistent water level series.