Retrospective problem (RR) was defined as systematic biases in successive estimations of fishery parameters (typically stock biomass and fishing mortality) with additional years of data added in stock assessment. RP arising in many fishery stock assessments can potentially result in large uncertainty in fish stock biomass estimates, which will lead to over exploit or under exploit fish populations. RP is one of hot and hard topics in fishery stock assessment at present. The MOHN ρ was used mainly as the metric of RP and the cause of RP was reduced to errors in data or in assumption of models or both. At present, there were diverse methods for diagnoses of RP, but these methods were generally valid only in some specific case. The researchers working on fisheries stock assessment provided some methods to fix or avoid RP according to their specific study, but there was still a lack of general methods or means to correct or avoid RP. If RP occurred in a stock assessment, there must be some errors in data or assumption of the model. Therefore, Retrospective analysis was effective means to assess the quality of the data and check the consistency of assumption between data and the model.