Pancreatic cancer is a highly malignant and metastatic cancer. Pancreatic cancer can lead to liver metastases, gallbladder metastases, and duodenum metastases. The identification of pancreatic cancer cells is essential for the diagnosis of metastatic cancer and exploration of carcinoma in situ. Organelles play an important role in maintaining the function of cells, the various cells show significant difference in organelle microenvironment. Herein, six probes were synthesized for targeting mitochondria, lysosomes, cell membranes, endoplasmic reticulum, Golgi apparatus, and lipid droplets. The six fluorescent probes formed OT‐SA (an organelles‐targeted sensor array) to image pancreatic metastatic cancer cells and cell spheroids. The homology of metastatic cancer cells brings the challenge for identification of these cells. The residual network (ResNet) model has been proven to automatically extract and select image features, which can figure out a subtle difference among the similar samples. Hence, OT‐SA was developed to identify pancreatic metastasis cells and cell spheroids combination with ResNet analysis. The identification accuracy for the pancreatic metastasis cells (> 99%) and pancreatic metastasis cell spheroids (> 99%) in test set was successfully achieved respectively. The organelles‐targeting sensor array provided a method for the identification of pancreatic cancer metastasis in cells and cell spheroids.This article is protected by copyright. All rights reserved