Indoor Visible light communication (VLC) is considered one of the most famous communication technologies in today’s industrial life, showing importance in data broadcasting and glowing instantly with the cost-effective source of light-emitting diode (LED). This kind of high-speed network’s properties is controlled by the source device’s shortened bandwidth (LED). Therefore, it is considering highly efficient modulation technique and extreme adaptive demodulation technique for better data rate in visible light communication. Carrier less amplitude-phase (CAP) modulation is an eminent modulation scheme that increases implementation ease and places good position inefficiency. Somehow, the CAP-VLC system’s impact is signal jamming, poor sensitivity, scattering, and noise issues. To overrule this problem, it witnesses the implementation of VLC system with CAP modulation using advanced neural network system of High-Speed Feed Forward Neural Network, which works based on the principle of Extreme Learning Machine (ELM). The same is adopted and extended using the BAT algorithm. Along with this, algorithm optimization technique has been integrated to enhance the entire CAP-VLC system’s performance. Altogether, it can be named as bat-optimized reliable ELM (BORE). The proposed new learning-based algorithm for CAP-VLC has shown better performance in received power distribution of 90.6% at various modulation indexes, better BER of about 97.6% in the voltage level of 3V, and different distances between transceivers, respectively.