Investor, consumer, and managerial decisions are central to both the causes and solutions to climate change. Yet the consequences of these decisions are often risky and complex, creating cognitive barriers to sustainable behavior. Building on recent theories of resource rationality and motivated cognition, we examine how information-processing constraints under risk and complexity shape sustainable decision-making. Across three incentive-compatible experiments, participants repeatedly chose between options differing in their climate and personal reward consequences. We manipulated each attribute to have either risky, complex, or certain outcomes. Sustainable behavior decreased when the climate consequences of one option were risky or complex, compared to when personal rewards were risky or complex, controlling for risk and complexity aversion effects. Moreover, complexity showed an asymmetric effect: when applied to climate consequences, it reduced sustainable choices relative to full certainty, but it had little effect when applied to personal reward. A calibrated multi-attribute decision model revealed that participants systematically placed less weight on climate consequences when they were risky or complex, consistent with resource rationality. Our findings demonstrate that sustainable preferences observed in simple and certain contexts may not generalize to real-world environments characterized by uncertainty and complexity. For managers and policymakers, the results suggest that simplifying climate information can promote sustainable investment, consumer, and organizational decisions.