NOMA technology has feactures such as high number of connections, spectral efficiency, low access delay and so on provides to the user. But this technology increases the complexity of the receiver in addition to the mentioned advantages. In general, NOMA works in two areas: coding and power. The field of coding itself is divided into subsets such as SCMA, LDS, LPMA, PDMA. The main activities carried out in the field of NOMA are in the field of reducing the complexity in the receiver and reducing the error rate in the receiver. In this thesis, we simulate coded LDS-NOMA system and investigate the effect of code length and repetition of information exchange between demodulator and decoder on error curves in Rayleigh channel. Here, in addition to using iterative detection in the demodulator, the exchange of iterative information between the decoder and the demodulator has also been used. Simulations show that repeating the information exchange between the demodulator and the decoder worsens the error curves. Also, increasing the length of the code also leads to the worsens of the error curves. The justification of this behavior can be justified in this way, because in the AWGN channel, the relationship between the probability of error and the distance is an exponential relationship, as a result, the error is greatly reduced by multiplying the distance. While in the Rayleigh channel, the relationship between the probability of error and the distance between two signals is fractional relationship, and with the multiplication of the distance, the process of reducing the error is less than that of the AWGN channel. Now, if we consider the coded system where there are several code words in the neighborhood of each code word, and let 𝑁 be the number of codes that are the minimum distance from the desired code. In this case, according to the error probability relations of code length and as a result of increasing 𝑁, the error rate decreasese in the AWGN channel, but the err