The paper is concerned with predicting the total cost of a new product and searching for cost reduction at the early stages of product development. The costs of a new product development project, product promotion, production and after-sales service are predicted using parametric models. The identified relationships are also used to searching for possibilities to reduce the cost of faulty products and after-sales service through increasing prototype tests. As a result, the trade-off between the cost of a product development project and costs of production and after-sales service are sought. Company resources and product specification are formulated in terms of variables and constraints that constitute the systems approach for a problem related to cost optimization. This problem is described in the form of a constraint satisfaction problem and implemented using constraint programming techniques. An example shows the applicability of the proposed approach in the context of searching for the desirable level of the cost related to prototype tests, faulty products and after-sales service. This study develops previous research in the context of adding the cost of after-sales service to a model of total costs of a new product. Moreover, the proposed method of predicting cost has been developed towards using the similarity value to data selection.