We reported results from a study on the effects of different training methods on complex perceptual-motor skill acquisition using a version of the Space Fortress game, which was originally designed to study the acquisition of complex perceptual-motor and cognitive skills in a multi-tasking environment. Participants were randomly assigned to the Fixed Priority (FP) and Varied Priority (VP) training conditions. Action sequences for controlling the spaceship in a frictionless environment using a joystick were analyzed and compared across conditions. Consistent with the previous findings, VP training was in general more successful than FP training. However, we found that VP training benefited participants more in the low performance group than in the high performance group. Participants in the VP training condition showed faster learning of optimal action sequences and faster reduction of suboptimal action sequences. In addition, results showed that in the high performance group, participants in the VP training condition used significantly more optimal action sequences than in the FP training condition. The findings have important implications on how the effectiveness of different training methods can be optimized for people with different cognitive abilities.