Fluorescence in situ hybridization (FISH) is a molecular cytogenetic technique that provides reliable imaging biomarkers for the diagnosis of cancer and genetic diseases at the cellular level. An important prerequisite for identifying cancer cells in FISH images is the accurate segmentation of cells to quantify the DNA/RNA signal within each cell. However, FISH images often have problems such as low contrast, unclear cell boundaries, and potential cell defects, which hinder the complete identification and effective segmentation of cells. Therefore, this paper proposes a cell image segmentation method for fluorescence in situ hybridization. First, an improved adaptive histogram equalization is used to enhance the image. Then, the watershed segmentation method is used to calculate the optimal segmentation feature value, which is used as the seed point of the watershed method. The proposed method was applied to segment 273 cells from 5 images of leukemia cells, with a detection rate of 99% and an effectiveness of 94.5%. The results showed that the fluorescence in situ hybridization cell image segmentation method can accurately segment cells, and has a high detection rate and effectiveness.