Currently, canine soft tissue sarcoma (STS) grading is based on histopathology. In humans, several studies have demonstrated concordance between cytologic grading systems for STS and histologic grade. The aim of this study was to correlate several cytologic parameters (smear cellularity, anisokaryosis, nucleolar malignancy score, multinucleation, and the number of mitotic figures per 200 cells) that form part of a human STS cytologic grading system, with histologic grades of canine cutaneous and subcutaneous STS. Three observers (blinded) reviewed the cytologic preparations independently from cases with confirmed histologic diagnoses of STS. A cytologic grading score was assigned for each parameter. Correlations between cytologic grading scores (averaged between observers) and histologic grades were assessed using Spearman's correlation coefficient, with statistical significance defined as P < .05. Twenty-one cases were included in the study (10 Grade I STS, nine Grade II STS, and two Grade III STS). The number of mitotic figures (≥3) per 200 cells was the only parameter that showed a significant but weak, positive correlation with histologic grade (rs  = .469; P = .032). No Grade I tumors had ≥3 mitotic figures per 200 cells; however, ≥3 mitotic figures per 200 cells were only observed in 33% of Grade II tumors and 50% (one out of two) of the Grade III tumors. This pilot study suggests that an increased number of mitotic figures seen on cytology might correlate with higher grade STS; however, the sensitivity of this parameter for grading STS appears to be low.