Abstract: Parameter fault and tolerance are two challenging problems in analog circuit fault diagnosis. This paper proposes approaches to solve such problems. First, a new fault modeling method and its theoretical proof are presented. In analog circuits, both the real part and the imaginary part of output voltage are the functions of fault component parameters. By eliminating component parameters from the simultaneous formulas, a new equation is achieved. It is independent from the value of component parameters and uniquely determined by the component locations and the topological structure. Hence, the function can be used as the fault model, which is applicable to both hard (open or short) and soft (parametric) faults. It is also applicable to either linear or nonlinear analog circuits. Then, the parameter tolerance is taken into consideration. A frequency selection method is proposed to maximize the difference between fault signatures. Hence, the aliasing problem arise from tolerance can be mitigated. The effectiveness of the proposed approaches is verified by simulated results.