The occurrence of lymph node metastasis (LNM) and depth of tumour infiltration are significant prognostic factors in oesophageal adenocarcinoma (OAC), however no reliable prognostic biomarkers have been established so far. Aim of this study was to characterize microRNAs (miRs) of OAC patients, who primarily underwent oesophagectomy, in order to identify specific alterations during tumour progression and LNM. MicroRNA array-based quantification analysis of 754 miRs, including tumour specimens of 12 patients with pT2 OAC from three different centres (detection group), was performed. We identified miR-17, miR-19a/b, miR-20a, and miR-106a, showing the best predictive power for LNM. These miRs were validated by quantitative real time-PCR (qRT-PCR) in 43 patients with different tumour stages (pT1: n = 21; pT2: n = 12 and pT3: n = 10) (training group) (p < 0.05), demonstrating that increasing levels of identified miRs were associated with advanced depth of tumour infiltration. These findings were verified in another independent group of 46 pT2 OAC patients (validation group). Quantitative RT-PCR analysis of the miR-panel confirmed these results except for miR-19a (p < 0.05 each). Logistic regression analysis identified miR-17 and miR-20a (p = 0.025 and p = 0.022, respectively) to be independent variables for prediction of LNM. The mathematical prediction model was used in the validation group, and the estimated prognosis was compared to the actual postsurgical follow-up. This comprehensive data demonstrated the importance of miR-17-92 cluster and miR-106a for progression as well as LNM in OAC indicating that those might be feasible prognostic biomarkers.