The dynamics of stock markets often exhibit complex behaviors that traditional models struggle to capture, particularly in recovery phases following periods of volatility. This paper introduces a novel mathematical model based on fractional calculus to describe and predict the rehabilitation dynamics of stock markets. We formulate a fractional differential equation (FDE) model and validate its effectiveness using historical market data. The model's ability to account for long-term memory effects and non-local interactions offers significant advantages over classical approaches