This study investigates whether statistical learning can tune attention to specific depth planes during visual search tasks in a virtual reality (VR) environment. Using VR to present a three-dimensional search space, we tested in two experiments whether the spatial distribution of targets and distractors across depth planes could be learned and utilized to improve search efficiency. The first experiment demonstrated that participants identified targets faster when they appeared more frequently on a specific depth plane, suggesting that statistical learning effectively guided attention allocation. The second experiment assessed the interference caused by distractors placed at varying depths. While the results did not reach statistical significance, there was a trend toward reduced distractor interference from less frequent and deeper depth planes. These findings indicate that attention allocation in 3D environments can be flexibly modulated by statistical regularities in depth, advancing our understanding of attentional control in complex perceptual spaces. VR offers a powerful tool for exploring these processes in immersive, ecologically valid scenarios.