It has been known that effects of visual attention are modulated by cue validity (i.e., proportion of valid and invalid condition in a certain block). The present study aimed at examining whether the benefits (i.e., the differences of reaction times in the valid and invalid condition) were linear as a function of cue reliability. We also examined effects of explicit instruction of information regarding cue reliability on visual attention. In the experiment, we adopted a cueing paradigm in which cue reliability varied in each block (0, 10, 20, 30, 40, 50, 60, 70, 80, 90, and 100%). We also classified all participants into two groups: Instruction and Non-instruction groups. The results showed that the benefit was linear as a function of cue reliability regardless of instruction, which indicates that cue validity has a strong influence on visual attention compared to a cued location. We also found that reaction time in the valid condition became shorter with increased cue reliability compared to the invalid condition with decreased cue reliability, albeit only in the Instruction group. This result indicates that participants with instruction focus more on the cued location (i.e., valid condition) and not the dominant proportion of cue validity.