In this paper, an approach of determining the secondary structures of RNA has been proposed based on the formalism of Constraint Satisfaction Problems (CSP). In this formalism, each base is associated to a variable while the experimental data and restrictions as well as the thermodynamic model are described in terms of constraints. In this formulation, filtering and searching algorithm within the CSP furnishes all the valid secondary structures of RNA. The use of a dynamic CSP allows interactive additions and relaxation of constraints whereas a weighted CSP offers a possible way of finding the sub-optimal secondary structures of RNA.