Global warming and climate change are detrimentally impacting the social environment and are increasing the death toll due to more severe natural disasters worldwide. For example, droughts in California, USA, caused economic losses amounting to US $2.2 billion and $3 billion in 2014 and 2015, respectively, and drought that occurred in India from 2000 to 2002 caused the deaths of 20 people and economic losses of US $1.5 billion. Besides, global arid regions are expected to increase ~ 10% by 2100 under a high-emission scenario. The more arid background witnesses the more frequent occurrence of severer, longer, and a really more extensive droughts that are likely to lead to water scarcity. Central northeast and west central India showed an increasing trend of the percentage (%) of the area under drought in the past (1950–2010), and % of the area under “above moderate drought” across India is expected to increase up to 70% by 2095 under the Representative Concentration Pathway (RCP) 8.5 scenario. Warming global temperature is also affecting the change in the location of deep convection over oceans and, in turn, the southwest monsoon rainfall, which contributes to 70–90% of the annual rainfall in India, tending to be dormant due to the amplified El Nino-Southern Oscillation (ENSO) and the stronger equatorial Indian Ocean positive anomalies. More frequent occurrence of Rossby waves, which stagnate atmospheric flow, and the intensified land-atmospheric feedback under warming climate make an already initiated drought more severe, longer, and spatially more extensive. Since drought occurs naturally in hydrologic processes, its occurrences are beyond human control. Thus, what humans can do is to be prepared for the anticipated droughts to mitigate drought impacts. To bolster relief efforts and mitigate drought impacts, a proactive drought management plan needs to be prepared, based on drought early warning, forecasting, and risk analysis. Application of deep learning, such as long short-term memory (LSTM) networks, and entropy spectral analysis have extended forecast lead time with higher accuracy than traditional statistical or machine learning models and have shown their applicability for proactive mitigation planning based on risk analysis. However, more improved monitoring systems for drought indicators, drought indices, which can consider implications of human activities and hydrologic processes explicitly, and more accurate drought forecasting is required to be prepared for severe droughts. Besides, historical drought impact data should be archived and organized for reliable assessment of vulnerability and risk to drought. Therefore, this study discusses the impacts of drought in the past and future and causal mechanisms based on drought events in India and suggests ways to be prepared for the expected more severe droughts from technical and governmental machinery perspectives.