In natural aquatic environments, the spectral composition of incident light often changes abruptly in response to the abundance of suspended particulate matter and zooplankton and plankton in nearshore waters. Understanding these responses is important for the aquaculture of various species, including the abalone Haliotis discus hannai. The current study investigated the metabolic regulatory mechanisms leading to growth differences of H. d. hannai under different spectral compositions [blue light (NB), orange light (NY), dark environment (ND) and natural light (NN)]. Overall, the cumulative distance moved, and duration and average speed of movement were significantly higher under NB than under ND or NY following an abrupt change from natural light to blue light. There was a significant increase in reactive oxygen species (ROS) content, with glutathione S-transferase (GST) activity and glutathione (GSH) content being significantly higher in Group NB than in Groups ND and NY, suggesting that the blue light environment causes oxidative stress to which the abalones respond by moving away. Based on LC-MS metabolomics, 201 and 105 metabolites were identified in NB versus NN, ND versus NN, and NY versus NN groups under positive and negative ion mode, respectively. Based on fold changes and annotations in the KEGG enrichment pathway, methionine, AMP, D-mannose 6-phosphate, phenylpyruvic acid, L-tyrosine, L-histidine, NAD+, and 5-hydroxytryptophan were used as metabolic biomarkers. Only Group NB had a significant increase in 5-HT concentration and 5-HT1A receptor expression levels, suggesting that this receptor is involved in the endocrine response of abalones to blue light stress. These results quantify the characteristics of the movement behavior of abalones under different light conditions, and preliminarily identify the physiological adaptation mechanisms of abalones under abrupt light change. Such data provide guidance for the optimization and regulation of environmental light factors in the aquaculture of H. d. hannai.