Settling particulate matter (SPM) was collected by using sediment traps at four stations in a survey section from Qingdao to Cheju-do, across the Huanghai Sea cold water mass (HSCWM), in August 2002.The sediment traps were planted in three layers:the upper layer of the thermocline (ULT) above the HSCWM, the lower layer of the thermocline (LLT), and the bottom layer of water column (BL).To determine the particle flux, the contents of organic carbon (POC), organic nitrogen (PON), total carbon (PC), and total phosphorous(PP) in SPM were analyzed, and two flux models (Ⅰ and Ⅱ) were improved to calculate the resuspension ratio, with an assumption in Model Ⅰ that the vertical flux of SPM in the LLT equals the net vertical flux of SPM in the whole water column.An X value, i.e., the fraction of the resuspension flux originating from the surficial sediments nearby the sampling station, was deduced from Model Ⅰ to estimate the contribution of lateral currents to the total resuspension flux.The results showed that inorganic particles, fecal pellets, and miscellaneous aggregates were the major types of SPM in the HSCWM, and the contents of POC, PON, PC, and PP all decreased with water depth.A great deal of fecal pellets found in the LLT indicates that the main space producing biogenic SPM is the thermocline, and especially the LLT, where the C/N ratio is lower than that in the ULT.The resuspension ratios, 90%~96% among stations, imply strong impact ofresuspension on particle flux in the BL.These values were not significantly different between the two flux models, suggesting that the hypothesis in Model Ⅰ that the flux in the LLT equaling the net flux to the bottom is acceptable for shallow waters with stratification like the HSCWM.The POC export ratio from the HSCWM ranges from 35% to 68%.It benefits from the short sinking distance in shallow water.The upwelling in the HSCWM enhanced the POC flux through the water mass, and the lateral currents provides up to being greater than 50% ofresuspension flux in the BL according to evaluation of the X value.