Comfortable walking speed (CWS) is indicative of clinically relevant factors in the elderly, such as fall risk and mortality. Standard CWS tests involve walking on a straight, unobstructed surface, while in reality surfaces are uneven and cluttered and so walkers rely on visually guided adaptations to avoid trips or slips. Hence, the predictive value of CWS may be expected to increase when assessed for walking in more realistic (visually guided) conditions. We examined CWS in young (n=18) and older (n=18) adults for both overground and treadmill walking. Overground CWS was assessed using the 10-meter walk test with and without visual stepping targets. For treadmill walking, four conditions were examined: (i) uncued walking, and (ii-iv) cued walking with visual stepping targets where the inter-stepping target distance varied by 0%, 20%, or 40%. Pre-experimental measures were taken so that the average inter-stepping target distance could be adjusted for each belt speed based on each participant's self-selected gait characteristics. Results showed that CWS was significantly slower when stepping targets were present in both overground (p<.001) and treadmill walking (p<.001). Thus, attuning steps to visual targets significantly affected CWS, even when the patterning of these targets matched the participant's own gait pattern (viz. 0%-treadmill-walking condition). Results from the treadmill-walking task showed that the amount of variation in inter-stepping target distance did not differentially affect CWS. Our results suggest that it may be worthwhile in clinical assessments to not only determine walking speed using standard conditions but also in situations that require visually guided stepping.