Channel estimation is crucial in wireless communication systems, as it intends to accurately determine the characteristics of the communication channel between the transmitter and the receiver. The accuracy of channel estimators significantly impacts the performance of various communication functions, such as beamforming and interference reduction. However, the traditional methods might struggle to cope with the dynamic and nonlinear nature of the wireless channels, especially in scenarios involving high mobility, dense multipath propagation, and frequency-selective fading. Recently, deep learning models, particularly convolutional neural networks (CNNs), have brought about significant advancements in wireless communication systems. In recent past, there has been a revolution in the channel estimation process, with CNNs demonstrating remarkable ability to capture the complex relationships between signals and channel conditions. Additionally, the long short-term memory (LSTM) models offer a compelling approach to processing even larger sets of input data and making them well suited for channel estimation tasks. This chapter delves into the modeling of optical wireless channels using both CNN and LSTM models, exploring their potential to enhance the accuracy and reliability of channel estimation in wireless communication systems.