Since the debut in 1988, 'Dangerous Dave' has been one of the most popular games. Playing game is great fun for every gamer. However, making agents play the game on their own and clearing levels matching human thinking is far better for software developers. Here we propose to leverage Deep Q learning and Actor-Critic to train agents and make them play on their own. The main objective of the bot is to explore the deserted pirate's hideout and collect golden trophies that are hidden by Dave's rival Clyde Cooper, who stole them from Dave's clubhouse.