The invention discloses a joint estimation method for the motion state of a vehicle and the gradient of a road. The joint estimation method comprises the following steps of constructing a ramp drivingplatform; estimating the motion state of the vehicle by using a square-root adaptive variational Bayesian cubature Kalman filtering algorithm, and forming a vehicle motion state estimation module; building a road gradient unit, and solving a road gradient value based on a ridge regression algorithm to form a road gradient estimation module; and adopting interaction iteration between the two modules to realize joint estimation of the vehicle motion state and the road gradient. The square-root adaptive variational Bayesian cubature Kalman filtering algorithm can adapt to the square root of a state covariance and the square root of a measurement noise covariance, and has the characteristics of small calculation amount, good real-time performance and high precision when compared with a similar noise adaptive algorithm. Based on the joint estimation for the vehicle motion state and the road gradient, the relationship between the vehicle motion state and the road gradient is considered, andthe accuracy of estimation results can be improved.