A method for estimating the component reliability is proposed when the probability density functions of stress and strength can not be exactly determined. For two groups of finite experimental data about the stress and strength,an interval statistics method is introduced. The processed results are formulated as two interval-valued random variables and are graphically represented by using two histograms. The lower and upper bounds of component reliability are proposed based on the universal generating function method and are calculated by solving two discrete stress-strength interference models. The graphical calculations of the proposed reliability bounds are presented through a numerical example and the confidence of the proposed reliability bounds is discussed to demonstrate the validity of the proposed method. It is showed that the proposed reliability bounds can undoubtedly bracket the real reliability value. The proposed method extends the exciting universal generating function method and can give an interval estimation of component reliability in the case of lake of sufficient experimental data. An application example is given to illustrate the proposed method. A method for estimating the component reliability is proposed when the probability density functions of stress and strength can not be precisely determined. For two groups of finite experimental data about the stress and strength, an interval statistics method is introduced. The lower and upper bounds of component reliability are based on the universal generating function method and are calculated by solving two discrete stress-strength interference models. The graphical calculations of the proposed reliability bounds are presented through a numerical example and the confidence of the proposed reliability bounds is discussed to demonstrate the validity of the proposed method. The proposed method extends the exciting universal generating function me thod and can give an interval estimation of component reliability in the case of lake of sufficient experimental data. An application example is given to illustrate the proposed method.