Purpose – The purpose of this paper is to separate households into several types based on their features, and then to further investigate determinants of household fish consumption in China by figuring out consumption preference divergences between types of households under the effects of economic and socio-demographics factors. Design/methodology/approach – This paper first applies Multiple Correspondence Analysis to separate the modalities of variables and households according to their features, with health knowledge and time constraint of a spouse highlighted. Then, the transcribed principal information of both variables and households has been added into Marshallian demand function with fish price, income, child effect, and health status for identification of factors on household fish demand. The robust fixed effect and robust random effect GLS regression has been conducted. Findings – The paper provides empirical insights about what and how factors affect household fish consumption. It suggests that,...