This paper proposes a method for remote sensing image band simulation based on spectral features study by SVM. As taking Spectral features into account as a priori knowledge and base on study of the implicit and nolinear relationships between landmark spectrum in the different obands by support vector machine, we simulated a new band through multi-spectral remote sensing image on existing bands. Based on TM blue band simulation experiments,We confirmed that this method can precisely simulate the real spectral images due to high reliability of the simulation results. Furthermore, Simulating IRS true-color experiments verify it is a practical method to reconstruct the missing bands of multi-spectral image. This method could theoretically possible to simulate a new remote sensing image band in arbitrary spectrometer's resolution within spectral range of landmark spectrum measure put forward a new idea for establishing links between landmark spectrum and remote sensing pixel spectrum.