Negative energy balance (NEB) is particularly prevalent in early lactation dairy cows and causes health problems that have a negative impact on farm economics. The gold standard for the direct diagnosis of NEB is the determination of serum non-esterified fatty acid (NEFA) concentrations, which is however quite costly and time consuming. Therefore, the aim of the study was to demonstrate the practical feasibility of estimating NEB from milk components using ROC analysis. A total of 193 dairy cows in the first 35 days in milk were included in the experiment. Blood and milk samples were taken from the cows on each milk yield recording day. Fat, protein and fatty acid (monounsaturated FA (MUFA) and C18:1) contents were determined from milk. NEFA concentrations were analysed from blood and dairy cows were divided into two groups: cows without NEB (NEB0; NEFA < 0.6 mmol/l) and cows with NEB (NEB+; NEFA > 0.6 mmol/l). Differences in milk composition between the groups were calculated and ROC analysis was performed to estimate NEB occurrence from milk components. During the first 35 days of lactation, 21% of dairy cows entered NEB. All milk components measured were significantly higher (P < 0.001) in the NEB+ group compared to the NEB0 group. Fat and protein contents, fat/protein ratio and milk fatty acids were included in the ROC analysis. Both MUFA and C18:1 were found to be excellent diagnostic markers for NEB (AUC GREATER-THAN OR EQUAL TO 0.8, Gini coefficient GREATER-THAN OR EQUAL TO 0.6). The ROC curves showed that if 80% of cows with NEB are detected using MUFA (i.e. 80% sensitivity), 29% of cases will also be false positives (i.e. specificity = 71%). In the case of C18:1, for 80% sensitivity, 70% specificity will apply, i.e. 30% of cows will be false positives. The cut-off values established for the detection of 80% of NEB were 1.35 g per 100 g of milk and higher for MUFA and 1.28 g per 100 g of milk and higher for C18:1.In conclusion, milk MUFA and C18:1 contents can be successfully used for early detection of NEB in the first 35 days of lactation. However, the cut-off values are valid for the herd under study and may be different in other conditions.