Abstract: The secondary ion mass spectrometry (SIMS) is the most adapted instrument for in situ oxygen isotope measurement of the coral samples. In situ analyses by SIMS have been limited because of the significant matrix effect, where the chemical composition and crystal structure affect the instrumental mass fractionation. The experiment was designed to evaluate the interference effect in this paper. The overall variability in IMS_coral in thickness variation was 1.02‰ (from -3.36‰ to -4.38‰), and about 0.7‰ in depth profile variation (from -2.75‰ to -3.45‰). In both cases, these variations could be related to changes in gas bubble on surface, time scale and vital effect. Changes in gas bubble may result in deformation of the electrostatic filed applied on the surface of the sample, which deviated the trajectory of ¹⁶O^- and ¹⁸O^-. It is clearly important to consider depth profile in estimating the ages of analyzed coral. The oxygen isotope was strongly influenced by the composition of the culture seawater and surface sea temperature (SST) that change over time. Therefore, understanding the analytical artifacts caused by sample preparation is significant for improving estimates of this coral matrix effect, which is a key part of SIMS and determines the veracity of data. Meanwhile, when the thickness was less than 3 mm and depth equal to IRMS sampling surface, IMS_coral was -2.75‰, which was comparable to the previous estimates of corals (IMS_coral=-2.8 ‰) within the error range. Therefore keeping the thickness of the coral sample less than 3 mm and depth equal to IRMS sampling surface is adequate to avoid interference on oxygen isotopes. This work guarantees the accuracy of in situ oxygen isotopes of coral and facilitates the SIMS application in coral geochemistry study.