Gastrointestinal (GI) cancers (gastric cancer, oesophageal cancer, liver cancer, colorectal cancer, etc.) are the most common cancers with the highest morbidity and mortality in the world. The therapy for most GI cancers is difficult and is associated with a poor prognosis. In China, upper GI cancers, mainly gastric cancer (GC) and oesophageal cancer (EC), are very common due to Chinese people’s characteristics, and more than half of patients are diagnosed with distant metastatic or locally advanced disease. Compared to other solid cancers, such as lung cancer and breast cancer, personalized therapies, especially targeted therapy and immunotherapy, in GC and EC are relatively lacking, leading to poor prognosis. For a long time, most studies were carried out by using in vitro cancer cell lines or in vivo cell line-derived xenograft models, which are unable to reproduce the characteristics of tumours derived from patients, leading to the possible misguidance of subsequent clinical validation. The patient-derived models represented by patient-derived organoid (PDO) and xenograft (PDX) models, known for their high preservation of patient tumour features, have emerged as a very popular platform that has been widely used in numerous studies, especially in the research and development of antitumour drugs and personalized medicine. Herein, based on some of the available published literature, we review the research and application status of PDO and PDX models in GC and EC, as well as detail their future challenges and prospects, to promote their use in basic and translational studies or personalized therapy.