Mars exploration missions increasingly demand reliable communication systems, yet harsh environmental conditions -- particularly frequent dust storms, extreme Doppler effects, and stringent resource constraints -- pose unprecedented challenges to conventional communication approaches. This paper presents the Martian Adaptive Sensing and Communication (MASC) system specifically designed for the Martian environment. MASC establishes a physically interpretable channel model and develops three key components: environment-aware hybrid precoding, adaptive parameter mapping, and robust communication precoding. Simulation results demonstrate that MASC maintains 45 percent sensing coverage under severe dust conditions compared to only 5 percent with conventional methods, provides up to 2.5 dB signal-to-interference-plus-noise ratio (SINR) improvement at 50 percent channel state information (CSI) uncertainty, and yields 80 percent higher capacity in moderate dust storms. Using an epsilon-constraint multi-objective optimization approach, we enable mission planners to select operational modes ranging from communication-priority (0.33 bps/Hz capacity, 28 percent sensing coverage) to sensing-priority (90 percent coverage with minimal capacity), offering a versatile framework that balances environmental awareness with hyper-reliable data transmission. This work provides a validated blueprint for integrated sensing and communication (ISAC) in non-terrestrial networks (NTN), a key enabler for achieving ubiquitous connectivity in the 6G era.