This study aimed to examine how the levels of physical fidelity of controllers (high, mid and low) and task repetitions (four trials) influence undergraduate students' perceived task load and performance in an immersive virtual reality (VR)-based simulation. The simulation was developed using the Unity 3D engine. VR controllers were developed to reflect a real power tool in three fidelity levels: high (most realistic: weight and tactile engagement), mid (tactile engagement without weight) and low (control group: only controller). The tasks were designed to reflect complexities with four working postures: no walking & standing up, no walking & bending over, walking & standing up and walking & bending over. Thirty-six healthy undergraduate male students participated in the study. Participants were instructed to complete motor tasks accurately. Audio feedback (drill sound) and haptic feedback (vibration) were activated for three groups when the controller and a screw made contact. Each participant used all three fidelity controllers and repeated the four tasks in a counterbalanced order to account for order effects. The results of a one-way repeated measures MANOVA indicated that two dimensions of task load were significantly different among the three physical fidelity conditions. Also, task completion time, inaccurate operation time and inaccurate counts were significantly shorter when four tasks were repeated. The study findings provide design implications for VR-based training experiences for future workforce development.