In this paper, the Genetic Algorithm based methodology for optimization of the apodization profile of a Super Structured Fiber Bragg Grating (SSFBG) coder, introduced in [1], is evaluated against the typical uniform profile used in Optical Code Division Multiplexing (OCDMA) applications. The maximization of codes cardinality as a function of the auto-correlation minimum (AC m ) and cross-correlation maximum (XC M ) levels for threshold detection at the decoder is analyzed. Inversely, for a given number of codes needed for a specific application, optimum threshold window can be chosen from a curve.This study demonstrates that given the same correlation constrains, for highly reflective SSFBGs (~4.5 dB loss per grating against typical 10 dB [2]), a large number of codes can be obtained by using optimal apodization encoders, whereas uniform profile encoders give only a residual numbers of well behaved codes. Therefore, the optimum apodization method enables low loss OCDMA based of fiber Bragg gratings. Nonlinear amplification is also discussed for detection threshold range conversion for the practical detection 7 dB standard level.