<p>The extravehicular material exposure experiments on the China Space Station (CSS) face a significant challenge: a mismatch between limited experimental resources and extensive scientific demands. This paper addresses the task planning problem of extravehicular material exposure experiments by constructing a unified modeling framework that covers experimental period, layout of exposure slots, and experiment scheduling. On this basis, a Mixed-Integer Linear Programming (MILP) model is formulated with the objective of maximizing the number of experimental projects, subject to primary constraints such as experimental period and the facility capacity. Furthermore, focusing on different engineering scheduling strategies, we develop and comparatively analyze two scheduling models: a fixed-frequency model and an optimizable-node model for payload installation and retrieval (PIR), examining their impact on experimental tasks. Numerical experiments demonstrate that the fixed-frequency PIR model, based on the current engineering practice, can provide decision-makers with selection recommendations among various options according to different planning priorities; however, it still faces a trade-off between the quantity and quality of experiments. Because fixed PIR nodes force actual exposure durations to passively adapt to engineering windows rather than strictly match scientifically required exposure periods, they may lead to an engineering-to-science constraint inversion in scheduling. To address this issue, the optimizable PIR-node model incorporates exposure-duration requirements as key constraints and optimizes PIR nodes while also considering the objective of maximizing the number of completed experiments. As a result, it can improve exposure-duration satisfaction while maintaining high experimental efficiency, thereby providing an effective decision-support method for task planning of extravehicular material exposure experiments on the space station.</p>