In order to overcome the major drawbacks existing in the simulation- based verification of floating- pointmultiplication,such as too large test space,dissatisfying test coverage and difficult corner case location,a genera-ting model of test case for normalized corner floating- point multiplication is constructed on the basis of Hensel lemma.Then,the test case and product are discussed concerning both positively-and negatively- normalized corner floating-point numbers.The results indicate that,in comparison with the traditional corner verification method,theproposed method improves the average error- detecting rate by about 7.39%; especially,it helps achieve an error-detecting rate at different floating- point bit widths up to 9.77%.Thus,it is concluded that the proposed modeleffectively improves the functional coverage and reliability of floating- point multiplication.