Applying different coding techniques to the Gaussian mixture traffic model. The effect of convolutional and Turbo coding on the probability of error and the capacity are investigated for CDMA IRIDIUM Low Earth Orbit Satellite (LEOS) systems. The model employed assumes a contaminated Gaussian traffic model. The conventional Gaussian distribution can be considered as a special case. The probability of error decreases using convolutional and turbo coding respectively. Furthermore, the capacity increases. Slight varitions around these values are obtained depending on the traffic nonuniformity. Thus Turbo coding improves the performance of the CDMA LEOS systems to a great extend.