The ecological risk pattern of the landscape on the Sanjiang Plain has experienced tremendous change in recent years, with the influence of natural processes and human activities. In this study, we collected data in the Sanjiang Plain from 1976 to 2013. Based on GIS and RS technologies, we established a landscape risk index, and simultaneously analyzed the corresponding driving forces for the ecological risk of landscapes in the Sanjiang Plain, using grid analysis and the geographic detector method. The results showed that landscape loss of farmland, grassland, and wetlands decreased slowly, whereas the loss of forest land and residential land continued to increase. Overall, the ecological risk was lower in the Sanjiang Plain from 1976-2013. From the spatial distribution perspective, the landscape ecosystem was primarily at a high risk in 1976, whereas it was primarily at a low risk in 2013. In addition, the areas with low, low-medium, and medium risks continued to expand into the vast west and southwest regions of the Sanjiang Plain, which was accompanied by a gradual retreat of the medium-high and high risk areas to the north of Heilongjiang and to the eastern Wusuli River along the protected areas and banks. From 1976 to 2013, ecological risk experienced a gradual migration to the southwest of the Sanjiang Plain, and the risk center of gravity migrated a distance of 37.8 km. The landscape pattern of ecological risk in the Sanjiang Plain experienced a change that was influence by altitude, geomorphology, residential distance, distance from the reserve, and human disturbance, whereas residential distance, human disturbance, and distance from the reserve explained much of the landscape ecological risk. The explanatory power of human disturbance (71.2%), residential distance (64.6%), and distance from the reserve (43.7%) was significantly higher than that of geomorphology type (11.2%) and basin type (8.9%) in 1995, and the explanatory power of the geomorphology type (23.5%) in 2013 was also slightly greater than that of distance from the reserve (23.3%). This result could provide a scientific basis for wetland protection and restoration.