Two experiments tested whether participants switched strategies while they are solving problems and age-related changes in such within-item strategy switching. Young and older adults performed a computational estimation task. Participants had to provide estimates of two-digit multiplication problems like 58 x 72, with either a rounding-down strategy (i.e., doing 50 x 70 = 3,500 to find a product estimate for 58 x 72) or a rounding-up strategy (i.e., doing 60 x 80 = 4,800). In Experiment 1, participants had the possibility to switch strategies after executing a current strategy for 1,000 ms if they judged the current strategy not the best strategy. In Experiment 2, participants were told to switch to another strategy 1,000 ms after starting to execute one strategy in strategy switch items, or to continue to execute the same strategy on strategy noswitch items. The main findings showed that (a) participants were able to switch strategies within items, especially when they started to execute the poorer strategy and when it was easier to determine which strategy is the best on a given problem, (b) older adults switched less often than young adults, especially in conditions where young adults switched most often, (c) switching helped participants to obtain estimates of higher precision, (d) switching incurred cognitive costs, especially when switching from a hard to an easy strategy, and (e) older adults' within-item strategy switch costs were larger than young adults', especially in conditions where participants tend to switch less often. These findings have important implications for furthering our understanding of strategy selection processes as well as of aging and strategic variations.