Channel estimation is a key feature for Free Space Optical (FSO) communication systems, necessary to ensure high-quality service and high data rates. Targeting the issue of classical FSO channel estimation, this work introduces a channel estimation scheme based on deep learning algorithms, namely Deep Neural Networks (DNN), Convolutional Neural Networks (CNN), and Recurrent Neural Networks (RNN). To compare with channel estimation techniques using classical algorithms such as Least Squares (LS) and Minimum Mean Square Error (MMSE), we apply these methods. Considering the presence of strong Gamma-Gamma atmospheric turbulence, we study the performance of the proposed structures. The results indicate that the proposed channel estimation schemes based on deep learning algorithms outperform traditional estimation techniques and can approach near-perfect channel estimation. Additionally, they are cost-effective, relatively simple, and offer favorable performance.