Light energy consumption minimization  has become  the important aspect of concern today . Here the Luminaires illuminance is controlled without the help of sensors. The approach is regarding feed forward neural networks implementation to model the relationships between the controller, luminaires and their luminance. The scheme is flexible to be implemented on microcontrollers unlike the conventional aproaches. This implementation has great modeling accuracy approximately by 95%. The energy saving with its optimal nonlinear multiple-input multiple-output control is also improved by more than 28%. Resulting in  ease of installation and cost effective alternate for reduction in lighting energy consumption.