Inflammatory skin disorder, eczema, is usually assessed by subjective disease scoring systems such as SCORAD and EASI. These scoring systems are based on clinical observations and questionnaires and hence it is subjected to inter and intra-assessor variability. Here, for the first time, we used optoacoustic imaging to image the structural and morphological changes of the skin in a non-invasive manner. Through a clinical study, we computed specific metrics such as epidermis thickness, total blood volume, vessel diameter in the dermis, ratio of low and high frequency signals. We trained a linear kernel-based support vector machine model for eczema classification using these metrics. We could achieve an accuracy of 86.6% and high sensitivity and specificity of 96.2% and 82.1% respectively. We also formulated a novel Eczema Vascular and Structural Index (EVSI) to objectively assess the severity of eczema.