Fractal dimension (FD) is a good option to quantitatively describe morphological complexity of cells. During the differentiation of rat bone marrow mesenchymal stem cells (rBMSC) into neural cells, morphological complexity of rBMSC increases, so FD can be a useful feature to characterize cell shape complexity at any given point in time during the differentiation process. In this study rat BMSCs were isolated and differentiated to neural cells. During the differentiation process, microscopic images were taken at certain intervals. After applying proper image processing algorithms implemented in Matlab software, FDs of the cell images were calculated by means of box counting method. Results demonstrated that FD of the recorded images had an ascending trend after one hour of neural induction, and simultaneously cells phenotype started changing from spindle shape to neural-like phenotype with multiple branches. Therefore, a direct correlation between cell shape complexity and FD of the images was observed, and it proves that FD serves as a practical method for quantitatively describing complexity as a morphological feature of the cells during differentiation.