As financial markets are volatile and rapidly changing, preciseness and agility in price evaluation and risk assessment in the portfolios are more important and decision support systems containing diverse financial products and pricing algorithms have been adopted to support those quantitative analysis tasks. To effectively use the financial DSS, the trader should be able to know which algorithms are applicable to a specific product and to match flexibly the product with appropriate algorithms depending on his product-evaluation purposes. This paper proposes model-base construction mechanisms in a financial DSS that facilitates such mix-and-match operations between financial products (considered as models) and pricing algorithms (considered as solvers). As a conceptual framework for the model-base, we use the generic model concept for developing system constructs and procedures. The constructs and procedures of the model-base are represented as self-contained and well-modularized objects that can be applied to a wide variety of problem domains.