The dynamic behavior of a cell model is affected by its structural complexity and parametric uncertainties. Two important issues in systems biology are how to quantitatively determine the relationship between system behaviors and parameter variations, and how to study the interactions between parameters. Using an NF-kB signaling pathway model as an example, and assuming that the parameters of this model are independent of each other and obey the identical uniform distribution in the range of variations, the global sensitivity analysis on the system output of NF-kB in the nucleus with respect to parameters is studied by means of the Latin hypercube sampling method. Simulation results demonstrate that the oscillation behavior of the concentration of NF-kB in the nucleus is sensitive to 6 key rate constants, which relates to reactions of IKKBa mRNA degradation, IkBa inducible mRNA synthesis, IKK adaption, constitutive IkBa mRNA translation, IKK-IkBa NF-kB association, and IkBbeta mRNA degradation, respectively.