The wheel speed sensor is comprised of two parts: rotary and fixed. Conventionally, the wheel speed sensor is mainly utilized to measure the vehicle velocity and acceleration. Because the rotary part and the fixed part are separate, the wheel speed varies with inducing factors, such as body pitching, suspension deflecting and wheel load fluctuating. Therefore, the wheel speed variation could be utilized to estimate vehicle attitude states. In this paper, correlations between the wheel speed variation and the inducing factors were formulated. The wheel speed could be described by a summation of four variations: body pitching, suspension deflecting, wheel load fluctuating, and driving torque. The summation correlation function was confirmed using actual vehicle tests. Based on the correlation function and a vehicle model, an estimation algorithm using a Kalman filter to calculate the body pitching velocity was proposed. In a test vehicle using the proposed algorithm, the estimation pitch velocity value and the measured value were in good agreement.