Abstract: Based on the yearly catch data (2001-2009) of Chinese fishery and inversion SST by satellite remote sensing in the Southeast Pacific Ocean, the predicting system on fishing grounds founded was developed based on the Bayesian probability theory. The system adopts the client/service mode; while the database runs under the SQL Server 2000 data management system combined the interactive controls GIS technology. The system was also verified on its precision by using historical catch data. The results showed that the prediction accuracy of fishing ground and non-fishing ground were 72.6% and 57.5%, respectively, while the average predict rate of accuracy was above 65%. The prediction accuracy of fishing ground in peak period of fishing was 3%-22% higher than which in fishing later period, but the prediction accuracy of non-fishing ground was 4%-11% lower than which in later period of fishing. Hence, the system has important advising significance in predicting fishing ground and fishing activity on the Chilean jack mackerel in the South Pacific Ocean. But in the actual predicting process, it needs to revise the forecasting location of fishing ground integrated other environmental factors, such as chlorophyll a, sea level height, ocean current.