We describe in this paper a Predictive Man—Machine Environment (PMME) that can in general be used in the automatic training and evaluation of control operators, and also serve as a Decision Support System (DSS) in their routine planning and decision making activity. In particular, we present the PMME in reference to Air Traffic Control (ATC), an economically and intellectually important domain for automation. The design and implementation of PMME are based on two connected workstations, one of which displays the essential features of the Current World and the other those of the Extrapolated World. The latter predicts the future consequences of the control operator's tentative decisions made in responding to the Current World. If the operator is satisfied with the consequences, he finalizes his decisions; otherwise, he modifies the tentative decision and, time permitting, goes through this cycle as many times as necessary. A simulated ATC environment for the PMME has also been implemented and two sets of experiments have been performed with tasks of equal complexity—one with and the other without the extrapolation facility. A statistically significant improvement has been shown in the ATC operations that rely on the extrapolation facility. It can be said in general that the PMME increases the effectiveness and efficiency of human judgmental processes by extending their range, capabilities and speed.