A novel algorithm for the position estimation of a mobile robot is proposed, which is based on probability statistics and grey system theory. For the proposed algorithm, the grey probability measure set is established, which is composed of a base set and the corresponding grey probability measure. The base set is used to represent the uncertain information and the grey probability measure distributes probability on the base set. Moreover, the integrating rules are formulated using the grey probability measure set and the q-satisfied rule to estimate the position of a mobile robot. In addition to providing a new way of representing the uncertain information, results of the proposed algorithm are also more reliable in the presence of error and outliers. The algorithm is applied in the position estimation of a Pioneer 3-DX robot in a corridor-office environment. Experimental results have shown that the estimation accuracy of the algorithm is as good as that of the particle filter and better than that of the extended Kalman filter.