The assessment of regime shifts and resilience variations are powerful analytical approaches for understanding the dynamics of social‐ecological systems (SESs) over time. Here, we used a sequential t‐test algorithm for analyzing regime shifts and system resilience coefficient (SRC) to assess the state of Lake Taihu, a shallow lake in China that has experienced long‐term external disturbances. The results showed that the Lake Taihu SES underwent a regime shift in the early of 1990s, and the development states of Lake Taihu SES can be divided into four stages: normal development stage (1960 to the late 1960s), preliminary degradation stage (the late 1960s to the late 1980s), disorder stage (the late1980s to the late 1990s), and reorganization stage (the late 1990s to the 2000s). The SRC declined gradually from 1960, reached its lowest point in 1990, and then slowly began to improve. Our results implied that the loss of system resilience resulted in the regime shift of the Lake Taihu SES essentially. Adaptive management that focuses on the gradual and fundamental variations should be adopted to prevent SESs from shifting to undesirable states.