Real-time information about vehicle mass and road slope is important for vehicle handling and stability control. This paper proposes a two-layer based adaptive parameter estimator to estimate the vehicle mass and road slope under longitudinal moving condition. In the first layer of the estimator, the slowly changed vehicle parameter, such as vehicle mass or road slope on the smooth road, is estimated by a computationally efficient adaptive law. Then in the second layer of the estimator, an online adaptive observer is applied to obtain the real-time information of the road slope on the uneven road. The proposed estimator is first verified by the numerical simulations. A scaled remote control car is then built to conduct the experimental tests. By using the measured data from the scaled vehicle, the proposed parameter estimator is validated to be able to estimate the vehicle mass and road slope in the real experimental conditions. In order to clearly present the estimation performance and compare estimation results, the root mean square (RMS) values and the normalised root mean square (NRMS) values of the estimation errors are calculated. The experimental results indicate that the obtained average RMS values of vehicle mass estimation error and road slope estimation error are 0.44 and 4.06, respectively. Road slope estimation shows bigger estimation error due to the severer vibration caused by climbing up the steep road.