<p>Financial time series forecasting is pivotal in various sectors such as portfolio management investment allocation and risk assessment. However, traditional methods are not suitable to provide acceptable forecasting attributes of financial data. In response, intelligent algorithms like artificial neural networks, fuzzy logic, machine and deep learning techniques have emerged as promising tools for improving forecasting accuracy. This article provides a comprehensive review of different intelligent forecasting models, highlighting their advantages and limitations. Directly feeding historical data into these models is not reliable due to the volatile, nonlinear, and non-stationary nature of financial information, along with uncertainties in numerous independent parameters. Therefore, additional feature engineering and optimal parameter selection techniques are necessary. Integrating signal processing techniques enables the extraction of useful features from noisy financial data, while optimization algorithms aid in model refinement and parameter tuning. The article also discusses performance indicators, which are crucial for identifying optimal features and parameters in forecasting models. These integrated approaches not only enhances forecasting accuracy but also provides new possibilities in decision-making processes across various sectors.</p>