Based on 22 sparse-distributed tide gauge records in the Northwest Pacific Ocean marginal sea, the process, characteristic and prediction of regional sea level change are discussed by the integration of the following methods. Firstly, the regularized EM algorithm (RegEM) and the Multi-taper Spectral Method (MTM) are adopted to interpret their multiscale fluctuation processes and their spatial-temporal variations. Secondly, the orderly cluster method is introduced to classify these tidal stations, and with the consideration of the space adjacent relation, we obtain five sub-regions (the coasts of Bohai Sea-northern Yellow Sea, Yellow Sea-East China Sea along Chinese coast, the East China Sea along Japanese coast, the southern East China Sea and the northwestern South China Sea). Furthermore, the Mean Generation Function (MGF) is explored to predict the medium- and long-term trends of each tide station. Finally, the Principal Component Analysis (PCA) is employed to obtain regional- scale sea level change trends, sea level rise rates of the above five sub-regions from 2001 to 2030 are 1.23–1.27 mm/a, 3.30–3.34 mm/a, 2.72–2.76 mm/a, 1.43–1.47 mm/a and 1.13–1.15 mm/a respectively, and the whole region sea level rise rate is between 2.01 mm/a and 2.11 mm/a. The aim of our work is to conduct an integrated research on regional sea level change.