Human balance creates stability and creates conditions for other movements to be carried out. Emerging from the interaction of several body systems, balance continually genera- tes small postural oscillations that may indicate health conditions. Postural oscillations have a complex geometric pattern and, for their analysis, fractal methods have been used. Fractals are fractured geometric structures composed of patterns that are repeated on at least two levels of scale. Fractal methods produce direct qualitative and quantitative cha- racteristics of movement and assess sensorimotor impairment. However, a large number of methods and the diversity of measurement instruments have been generating different interpretations in the clinical area, making it difficult to create reference values. This work aimed to propose and validate a protocol for fractal analysis of human balance. Considering postural oscillation as a bivariate distribution, distributed in the anterior- posterior and medial-lateral coordinates, the protocol developed is an optical-electronic system. To capture data in these two coordinates, the Computational Vision Mobility software version 3.6 was used and the box count dimension method with the algorithm described by Higuchi to estimate the fractal dimension of the postural oscillation in the considered directions. The protocol was validated by investigating the postural oscilla- tion of 45 individuals, 23 young adults and 22 mature adults, and comparing the results obtained with the results of authors who used the box count dimension method, especi- ally the Higuchi algorithm. Thus, the proposed protocol is an objective analysis system of postural oscillation in healthy individuals that provides reliable and more in-depth in- formation on postural oscillations. Also, the protocol is portable, non-intrusive, occupies little space, and is easy to operate, with different ways of assembling the components allowing a better adaptation to the installed location.