Abstract: To solve the high complexity problem of OBB intersect test during collision detection process in virtual environments, this paper proposes a fast oriented bounding box（OBB） intersect test algorithm which is based on linear inequalities effective constraints. First, we describe OBB by linear inequalities, and give the associated mathematical theory conditions in of matrix dimensionality reduction, i.e. effective constraint theorem. Second, we simplify the original coefficient matrix via utilizing the corresponding row of the hyperplane which corresponds to effective constraint inequality, namely achieved matrix dimension reduction by excluding related rows and columns when necessary by operating the first column of the matrix elements using elementary transformation. We also make the comparisons about the performance and accuracy in static case and dynamic case. Finally, the accuracy and time complexity of the algorithm is briefly analyzed. The experimental results show that the performance of the algorithm is faster than traditional algorithm in 2~3 times, and the method has the overall superiority.