The number of available walking tests has increased dramatically over the past decades. Therefore, it is highly important to help clinicians choose the most appropriate walking test for a specific setting. This systematic review aimed to critically evaluate the reliability, validity, and responsiveness of clinical walking speed in a broad population of elderly persons living in the community, sheltered housing, or institutions. Literature searches were performed in several different databases. Key words were based on the topic of the measurement properties of performance-based clinical tools for quantifying walking. The instrument selected for review was walking speed. The methods and results of all the evaluated measurement properties were rated by using a standard checklist for appraising the qualitative attributes and measurement properties of the instrument. A total number of 3,781 abstracts were reviewed, and 86 articles were chosen for inclusion. Habitual walking speed seems to be highly reliable in community-dwelling people and residents in mixed settings. There have not been any studies that accord with our inclusion and exclusion criteria that have evaluated the reliability of maximum walking speed in an aged population. Walking speed is a highly valid test, both at habitual and maximum speed. Few studies gave information about responsiveness for walking speed, which means that these results cannot be evaluated properly. Habitual walking speed is a reliable measure, but maximum walking speed needs further evaluation. Both habitual and maximum walking speeds are valid instruments, and they predict death, hospitalization/institutionalization, and decline in mobility.