This study investigated the effect of time, gender, and direct strategy instructions under unlimited time condition on Vandenberg and Kuse revised version of Mental Rotation Test (VMRT). 129 college students in University of Macau were employed and randomly assigned into three treatment groups (holistic, analytic, and neutral strategy instruction) and took two separated tests (VMRT). The results indicated that the male outperformed than female with a small effect size (partial η2 = .062). Though the result showed a significant effect of time (pretest, posttest), the strategy instruction had no significant effect. Thus, the time effect came from the practice. Additionally, females appeared to guess more in untimed limit condition, and with more practice, participants became more confidence and accurate.