The remarkable expansion of wireless communication systems has given rise to the development of all communication generations, spanning from 1G to 5G/6G. Nevertheless, the ongoing focus is on the research and development of 6G wireless and mobile communication systems according to established timelines, and the primary goal is to cater to the substantial demands for data traffic. Evidently, the concept of “connected intelligence” serves as the foundational principle for the upcoming era of wireless communication systems. Meeting the demands for massive connectivity, particularly in the high unlicensed bandwidth, poses a challenge for conventional RF systems and spectrum alone. To address this, the millimeter wave (mmWave) range has been considered as an alternative. However, the reliance on mmWave communication may introduce additional regulatory constraints, potentially falling short of fulfilling the intended purpose. Consequently, the optical regime of the spectrum, particularly FSO communication systems, has emerged as a significant contender. Operating in the infrared wavelength of the spectrum, FSO communication systems enable operators to deliver bandwidth capable of supporting trillions of connected wireless devices. FSO systems possess properties that make them well suited for meeting the demanding requirements of secure, ultrareliable, low-latency communication in 6G systems. Additionally, addressing the needs of self-configuration, context awareness, and cellular aggregation in 6G systems can be accomplished through the integration of AI, particularly leveraging deep learning models. The inherent capacity of deep learning models to comprehend datasets proves valuable in facilitating device-to-device communication and configuring heterogeneous networks within the 6G framework. The combination of deep learning and FSO systems demonstrates the capability to fulfill the stringent requirements of the next generation of wireless and mobile communication systems. The present chapter focuses on implementing channel estimation method for FSO systems. The point of discussion in this chapter is in accordance with the various problems that are inherent in the modeling of the optical wireless channel. The research attempts to use the high-dimension feature of the deep learning models to represent the complex varying FSO channel.