The increased use of renewable energy sources has brought numerous environmental benefits. However, a significant challenge with renewable energy is the inability to control production, leading to over or under-production, both with considerable consequences. To maintain grid balance, energy companies can participate in ancillary service markets, offering their services through bidding systems. Accurately predicting these market trends can be highly profitable but is complex due to the variability in prices and volumes. Solving this can help energy companies utilize renewable energy sources on a greater scale to increase profit, which would greatly benefit the environment. This thesis investigates the optimal technology for forecasting these markets, assesses their accuracy, and explores how to use these predictions for effective bidding strategies. The research uses a quantitative approach with an experimental design to evaluate various machine learning models for predicting the Swedish ancillary service market. It incorporates insights from the Swedish energy company Mälarenergi, which are used in combination with the predictions to create a bidding strategy tool. The findings suggest that while some services are harder to predict, the current optimal technologies for forecasting are identified. Additionally, the need for dynamic methods in determining optimal bidding strategies is highlighted. The thesis concludes by emphasizing the necessity for ongoing research due to the evolving nature of energy markets and their dynamics.