This is a follow-up study of Chen's validity study in ATC selection in 1993. A total of 66 ATCs were included. Because a standard evaluation system of instructors has not well built up, the criterion of ratings from instructors was not included in the validity analysis. Only the psychological aptitude and psychomotor tests were used to predict the training performance. The criterion of training performance included curricula scores, on-the-job training, general act evaluation, and total scores. The results showed that some tests having low correlation with criterion were speed of attention, speed of spatial memory, speed of choice reaction time, depth perception, mean and stability of high speed anticipation, and mean of low speed anticipation. Generally speaking, aptitude and psychomotor tests both could predict 73% variances of training performance.