With the technology advancement, electronic devices are miniaturized at every development node. Physical parameters such as microscopic roughness are affecting these devices because surface to volume ratio is increasing rapidly. On all the real surfaces microscopic roughness appears, which affects many electronic properties of the material, which in turn decides the yield and reliability of the devices. Different type of parameters and simulation methods are used to describe the surface roughness. Classically surface roughness was modeled by methods such as power series and Fast Fourier Transform (FFT). Limitations of this methods lead to use the concept of self-similar fractals to model the rough surface through Mandelbrot-Weierstrass function. It is difficult to express surface roughness as a function of process parameters in the form of analytical functions. Method based on neural networks has been used to model these surfaces to map the process parameters to roughness parameters. Finally. change in electrical parameters such as capacitance, resistance and noise due to surface roughness has been computed by numerical methods. (C) 2004 Elsevier B.V. All rights reserved.