Diagnosing wooden foreign bodies (WFBs) using computed tomography (CT) is often missed, leading to adverse outcomes. This study aims to reduce misdiagnoses by exploring the density variation of blood-saline mixtures in ex vivo models. Twenty Cunninghamia lanceolata sticks, selected as WFB models, were randomly assigned to five groups: a control group (saline) and four experimental groups immersed in blood-saline mixtures with varying concentrations. The samples were then placed in a constant-temperature water bath at 36.8 °C. CT scans were performed in the lowest and highest density areas, and the volume of the low-density areas was measured at the post-processing workstation. Finally, the effects of time and concentration on imaging were analyzed, and fitting curves were generated. The blood-saline mixture concentration and time significantly affected the CT number in the three areas. WFB images changed dynamically over time, with two typical imaging signs: the bull's-eye sign on the short axis images and the tram line sign on the long axis images. Fitting curves of the CT number in the lowest density areas with different concentrations can quantify imaging changes. The CT number of the lowest density areas increased with time, following a logarithmic function type, while the CT number of the highest density areas exhibited a fast-rising platform type. The volume of the low-density areas decreased over time. The time of damage caused by WFBs and the influence of varying blood and tissue fluid contents at the damaged site should be considered in the diagnosis. Imaging changes from multiple CT scans at different times can aid in diagnosis.