Abstract: Aiming at the information classification and attribute reduction between multiple universes in the real world, a U×W double-disciplinary variable precision rough set model is proposed. By studying the equivalence relation in the U×W double domain, Based on the equivalence class of two-dimensional equivalence relation, the structure and mathematical characteristics of the model are deeply studied, and some useful conclusions are obtained. Finally, the uncertainty is measured by rough entropy and information entropy. The model and measurement results are verified by examples.