Geo-event is controlled by the environment factors with nonlinear relationship. That means it is important to discover the spatiotemporal assignment of environmental factors. Vector-based association rule discovery models have been provided to look for environmental pattern successfully in terrestrial applications recently. But they just consider the topological relationship in static state between parcels or objects with complex algorithm. And the data type of the attribution has to be limited as category. Except the topologic relationship

it is difficult to take other relationships into account. In the environment research

the continue field also is important research content

especially in meteorology and oceanography. And the continue field always includes the temporal property as geographic phenomenon or geographic progress. This work is to deal with the process. It disperses the space with time. A neighborhood is defined to extract the spatiotemporal variables to make a decision table. A reduction algorithm based on rough set will mine the spatiotemporal assignment with the environmental factors. Finally

the model was applied in fishery geography to find the environmental pattern

which determines the form of fishing ground.