Economic impact of subclinical ketosis (SCK) in dairy herds resulted from a decreased productivity, reproductive performances, and health condition. In Algeria, data availability regarding the prevalence of SCK in dairy cows were limited. The present study is a contribution in that direction and aimed to determine the prevalence of SCK between 2 and 50 days postpartum whereas investigating herd-and cow-level risk factors and the metabolic profile associated, on two bovine populations imported and native. The study was conducted with 16 typical representative local dairy cow herds (n = 100; 2-12 years old; located at Merouana, Batna Province) with data collected in a period comprised between August and February of the following year. Concentration of cows' plasma beta-hydroxybutyrate (BHBA) was used for the diagnosis, with values equal or higher than 1.2 mmol/L chosen as the threshold for SCK detection. Other key associated zootechnical parameters such as body condition score (BCS) were monitored at the same sampling points. Logistic regression analyses were carried out to characterize association between parameters. The average SCK prevalence was 9% and resulted highest during first-and second-week postcalving, with a peak in the first week. The herd size (< 30 lactating cows) and BCS (≤ 3) were risk factors associated with SCK prevalence. At the cow-level, risks of SCK increased in thin individuals (P = 0.058; odds ratio (OR) = 5.00, 95% confidence interval (CI) = 0.95-26.32). At the herd-level, risks of SCK increased in smaller frame size herds (P = 0.047; OR = 8.76, 95% CI = 1.03-74.49). Most cases of SCK were observed in native cows compared to imported. Interestingly, linear stepwise regression showed a good relationship between BHBA and total bilirubin (P = 0.0001; R 2 = 0.59). There were significant differences between healthy and ketotic cows regarding plasma concentration of bilirubin (P = 0.003), glucose (P = 0.001), alanine aminotransferase (ALT) (P = 0.008), creatinine (P = 0.007), and sodium (P = 0.02). According to receiver operating characteristic (ROC) curve analysis, the best parameter for SCK prediction is total bilirubin with more than 4.09 g/L as a cutoff point (77.78% sensitivity and 94.51% specificity). According to metabolic profile and feeding data we suggest that most cases of SCK observed in our study are type I. Further studies are required to validate other metabolic predictors for SCK and investigate adaptation of imported cows in semiarid regions in Algeria.