In order to solve the problem that the number of classification boundaries of convex hull region formed by polyhedral cone classifier is limited and it is unable to scale in different scales，a rotated and extended polyhedral conic classifier（REPCC）is proposed by adding a rotation factor based on norm vectorization. REPCC in? creases the number of classification boundaries of the convex hull region，and the classification boundaries can be adaptively scaled in each dimension，which can better fit the positive region and improve the classification accuracy. Experimental verification is carried out on two different rolling bearing datasets. The results show that REPCC has better classification accuracy，robustness and generalization ability，and can accurately identify the working state and fault type of rolling bearing. REPCC can be used for intelligent fault diagnosis of rolling bearing.