Current methods of drought assessment utilize drought indices, such as the standardized precipitation index and Palmer drought severity index, that rely on subjective thresholds and hence cannot be universally applied across different climatic regions. In addition, most of the existing drought indices are not amenable to probabilistic treatment which is essential for quantifying model uncertainties in drought classification. This study applies a machine learning tool, the hidden Markov model (HMM), for probabilistic drought classification. The HMM-based drought index (HMM-DI) developed in this study, does not require specification of subjective thresholds and model parameters are determined from historical data during parameter estimation. The drought classifications obtained using HMM-DI are compared with SPI results. The HMM-DI reveals new insights into the frequency and severity of droughts and their spatio-temporal variations. The effectiveness of HMM-DI is assessed by its application to monthly precipitation data over India. The results suggest that HMM-DI can be a promising alternative to conventional drought indices.