Abstract: This study examined spatial-temporal variations of water quality and potential pollutant sources in the surface water layer in the offshore areas of Bohai Bay, China, based on 7,488 sets of experimental data collected for 13 water quality indicators at 12 sites three times yearly from 1995 to 2010. A multivariate statistical approach was comprehensively applied, including cluster analysis, discriminant analysis, principal component analysis, and factor analysis. The data series of 16 years were grouped into three clusters: cluster A at early stage of nutrient pollution dominant (1995-2001 with 1998-1999 excluded), cluster B at middle stage of organic contamination dominant (2007-2008), and cluster C at late stage of land-based pollution and human activities dominant (1998-1999, 2002-2006 and 2009-2010). In spatial variation, human activities had more serious impact on the nearshore stations; oil and nutrients pollution was more serious in the marine culture areas. This study shows that the comprehensive approach of multivariate statistical methods is effective and useful in water quality investigation of the Bohai Bay offshore areas and our findings would help further study of water ecological environment in this area.