The aim of the study was to determine whether the addition of retinal or extra-retinal information could be used to improve memory-guided saccadic performance in healthy participants. Furthermore, we included two age groups in our study; healthy young adult subjects (mean age 20 years) and healthy middle-aged adult subjects (mean age 52 years). All subjects performed a novel task that incorporated a Go/NoGo task with a memory-guided saccade paradigm to investigate whether extra-retinal information (making a saccade towards the visible target) or retinal (a visible frame-of-reference) has any affect on the accuracy or variability of the response. We found all subjects made slight hypometric responses to the memory-guided targets. Both younger and middle-aged subjects revealed an increase in accuracy in the Go task compared with the NoGo task and the framed condition compared to the frameless condition, respectively. The frame also revealed a significant decrease in variability in the memory-guided saccades. A positive correlation in errors between the 1st and 2nd saccade in the Go task was revealed for all subjects; however, the older subjects revealed a greater correlation than younger subjects. The results presented indicate that younger and middle-aged perform highly similar patterns of errors during eye movements to remembered locations. However, middle-aged subjects show a greater tendency to use extra-retinal and retinal feedback to guide the response.