Introduction: The Exeter short stem (ESS) is 25 mm shorter than the standard length v40 Exeter stem (Stryker) and intended for a narrow femoral diaphysis. The purpose of the study was to evaluate the migration pattern of the cemented ESS. Material and method: In a prospective single-center cohort study, 23 patients (21 female) mean age 78 (range 70–89) with hip osteoarthritis and Dorr Type A femurs were included. Preoperative DXA was used to group patients into normal (> − 1) and low (< − 1) T-score. Components were the collarless polished double-tapered Exeter short stem type N°1 L125. Patients were followed for 2 years with model-based RSA (stem migration), regular hip radiographs (stem position and cementation quality), Oxford Hip Score (OHS) and VAS pain. Results: At 2-year follow-up, the stems subsided 1.48 mm (CI 95% 1.69; 1.26) and retroverted 0.45° (CI 95% 0.01; 0.88). From 12 to 24 months, stem subsidence was 0.18 mm (CI 95% 0.1; 0.25) (p = 0.001) and retroversion was − 0.04° (CI 95% − 0.27; 0.18) (p = 0.70). T-score and stem subsidence correlated (rho = 0.48; p = 0.025) and patients with normal T-score (n = 7) had 0.42 mm (CI 95% − 0.01; 0.85) less subsidence as compared to patients with low T-score (n = 15) (p = 0.054). Stems in varus position (n = 9) subsided 1.7 mm (CI 95% 1.35; 2.05) compared to 1.33 mm (CI 95% 1.05; 1.60) for stems in neutral position (n = 13) (p = 0.07). Postoperative cementation quality did not influence 2-year stem migration. OHS improved to 40.7 (CI 95% 36.8; 44.7) and VAS pain at rest and activity decreased to 5 mm and 10 mm, respectively (p < 0.001). Conclusion: The 2-year migration pattern of the cemented ESS was similar to reports for the cemented standard length Exeter stem. Low preoperative T-score and varus stem-position showed a tendency for higher stem migration and should be studied as risk factors for failure in larger studies of cemented polished stems.