In the application to nanometre resolution lithography of the metallic near-field superlens, the image quality becomes a
critical issue. Fundamental Fourier optics is applied to analyze the image system. The transfer function is computed
with the transfer matrix method, the Surface Plasmon Polariton (SPP) resonance and the SPP waveguide theory.
However, as the scattering of the object nano-structure involving the solution of the Maxwell's equations, so that the
object function is in general unknown, and the impulse response is less likely useful for computing the image.
Especially, metal object may induce the electrical dipoles, which launch the SPP and act as sources of radiation. The
superlens may be optimized based on the transfer function using the long-range SPP mode cut-off technique, the genetic
algorithm and other techniques in order to improve significantly the image quality. Design examples are presented, and
confirmed by the real image computed with numerical simulation using the FDTD method.