Objective: The characterization of stroke lesions is a challenging research issue due to the wide variability in the structure of lesion patterns. The objective of this research work is to characterize the stroke lesion structures using fractal analysis.Methods: To characterize the complex nature of the lesion structures, fractal box counting analysis is presented in this work. Three parameters from fractal dimension (FD) are considered to characterize the nature of the normal and abnormal brain tissues.Results: The experimental results are presented for 15 different datasets. Three different parameters namely FD average, FD deviation, and FDlacunarity are extracted to quantify the properties of the stroke lesion. The observations indicate that there is a significant proportion of separationof feature values between the normal and abnormal brain tissues.Conclusion: This work presents an efficient scheme for characterizing the stroke lesions using fractal parameters. It could be further enhanced by incorporating features extracted from other non-linear techniques.