This paper presents the results of tests conducted with an ultrasonic proximity measurement system (composed of Polaroid 600 sensors and a Sonar Ranging Module SN28827), which is used in sensory subsystems in many mobile robots. The tests took into account distances ranging from 0.4 to 11 m. The analysis of the obtained measuring results created the basis for selecting a distance estimator that would guarantee the smallest measuring errors in the whole measuring range of the system. The research described was carried out in a closed room at constant environmental conditions so as to eliminate the influence of external factors on the measurement result. Each measuring series was composed of 100 measurements. For the best estimator chosen in such environment, the value of absolute measuring error never exceeded plusmn 0.03 m. Histograms that present the scatter of measurement results are also included. The minimal number of measurements necessary to achieve a reliable measurement with the selected distance estimator was determined. It is shown that in order to achieve a relative measuring error smaller than 1% with 0.99 probability there is a need to perform at least seven measurements. A proposition of the distance measuring procedure with the chosen estimator is presented. The analysis described in this paper helps to evaluate the reliability of measurements performed with ultrasonic sensors.