INTRODUCTION: Poincare plots remain largely unused for heart rate variability (HRV) analysis in dogs. The aims of this study were to describe, qualitatively and quantitatively, Poincare plots in healthy dogs, to compare them with other methods of HRV analysis, to assess their day-to-day variability and to investigate the effect of activity on the plots. ANIMALS: Twenty-five healthy dogs. METHODS: Poincare plots, their standard descriptors, and other measures of HRV were generated from 24-hour ambulatory electrocardiographic recordings and 6 hours of rest and activity. RESULTS: The 24-hour Poincare plot demonstrated a \'Y\' pattern. The arms of the \'Y\' were derived mostly from periods of rest and activity populated the stalk. The quantitative descriptors of the plot had strong correlations in the time-domain, with weaker correlations in the frequency-domain. Individuals showed low day-to-day variability of the plot pattern and of the standard deviation of points along the major axis of the plot (SD2), which measures overall HRV. Day-to-day variability was higher for the standard deviation of points perpendicular to the major axis of the plot (SD1), which is a measure of short-term HRV, and for the SD1/SD2 ratio. CONCLUSIONS: Twenty-four-hour Poincare plots in healthy dogs show a \'Y\' pattern with subtle variations unique to the individual. The amount of activity and rest within the recording has a significant effect on the plot. Quantitative analysis of the plot can be used as a surrogate for time-domain analysis of HRV but visual analysis of the pattern provides additional information.