The foreign exchange (forex) market, characterized by its high volatility and complexity, presents significant challenges for accurate prediction. Traditional methods, including technical and fundamental analysis have been extensively utilized; but often fall short in capturing the multifaceted dynamics of the market. This paper provides a comprehensive literature survey on forex prediction, focusing on a hybrid approach that integrates technical indicators with other predictive techniques. We begin by outlining the key challenges in forex prediction, followed by an overview of traditional methods, highlighting their strengths and limitations. This survey centers on a comparative evaluation of seven recent studies on forex prediction, each employing diverse techniques, ranging from conventional machine learning algorithms to advanced deep learning methods. A comparison table summarizes these studies, detailing their methodologies, advantages, and limitations. This survey aims to identify gaps in current literature and explore the capability of hybrid models to improve the accuracy of forex predictions by leveraging the complementary strengths of various techniques. The findings imply that while conventional methods remain valuable, their integration with advanced computational techniques offers a promising avenue for improving predictive performance in the forex market.