Background: Few reported studies describe normal values from 24‐hour ECG (Holter) recordings of small breed dogs. Objectives: To investigate influence of breed, age, sex, body weight, degree of recording artifact, and mitral valve prolapse (MVP) on Holter recordings of 3 breeds of small dogs that have differing predispositions for myxomatous mitral valve disease. The study also assessed if heart rate (HR) at clinical examination (HRex) was associated with HR during Holter monitoring and evaluated the reproducibility of Holter variables. Animals: Fifty clinically healthy, privately owned dogs of the breeds Cavalier King Charles Spaniel (CKCS), Wire‐haired Dachshund (wD), or Cairn Terrier (CT). Methods: Prospective, longitudinal observational study. Dogs were recruited for clinical examination, echocardiography, and Holter monitoring. In 8 CKCS, Holter recordings were performed twice with a 7‐day interval. Arrhythmia and heart rate variability (HRV) analysis (time and frequency domain analysis) were performed on Holter recordings. Results: Fifteen out of 27 Holter derived variables were significantly associated with breed (P < .03), but not with age (P > .7), sex (P > .2), body weight (P > .7), degree of recording artifact (P > .4), or MVP (P > .6). During Holter recording, minimum (P= .0001) and mean HR (P= .0001) were higher in CKCS compared with wD. CKCS had significantly lower values than wD, CT, or both in 10 out of 13 HRV variables (P < .03). Minimum and mean HR during Holter recording were correlated with HRex (r= 0.55, P= .0003). HR and time domain variables had a coefficient of variation <10%. Conclusions and Clinical Importance: There is an influence of breed on Holter‐derived variables in 3 breeds of small dogs. Arrhythmia and HRV analysis can be performed on 24‐hour ambulatory ECG (Holter) recordings. Arrhythmia analysis includes HR measurements and identification of arrhythmias.