This research paper proposes an approach to detect and correct concept drift in time series based financial forecasting information. The test-then-train methodology is used to handle non-stationary streaming data, and the proposed system detects drifts in data using Autoregressive Integrated Moving Average (ARIMA), Seasonal ARIMA (SARIMA), and Hellinger Distance Drift Magnitude (HDDM). The approach seeks to make better and more flexible predictions in volatile financial markets since it keeps retraining the model as new concept drifts are detected, thus helping to provide more accurate asset price prediction. The technique is shown to be effective experimentally on historical stock price data, with active retraining of the models increasing forecast accuracy in the face of concept drift. SARIMA+HDDM outperforms among the studied models in forecasting dynamic financial data.