Recent findings from behavioral finance indicate that cognitive support is critical to investors as their psychological biases strongly influence their investment decisions. This paper proposes a conceptual model of investor misjudgment based on the three-stage human information processing model. The importance of such a model is that it classifies investment-related biases as being long-term or short-tern and consequently, provides a way to implement debiasing mechanisms in DSS. The paper then suggests an architecture for building such a cognitive investment DSS using recent computational technologies for human attitudes modeling. This research work fills a gap in the current IS literature related to behavioral finance and offers a novel approach for integrating findings from that domain into a cognitive investment DSS.