Sensor network-based home automation systems are familiar over the recent decades. Incorporating the benefits of the sensor network, energy management systems (EMS), is introduced to benefit end-user through periodic information sharing and remote access. WSN opted for energy harvesters to reduce the maintenance costs and maximize the lifetime of network. It is a perfect match for wireless devices and WSNs. Energy management system designed for effective use of harvested energy. Wireless sensor networks (WSN) coupled with EMS and grid-based applications serve as a support for smart home appliances. The integrated system architectures are cost effective and are energy harvesting that is profitable for end-user applications. Identifying optimal devices and defining an energy management policy are a tedious task as the devices are interfaced through different application support. This manuscript proposes a reward-based energy harvesting (REH) approach for identifying reliable devices in order to frame minimal-allocation energy for its operation. The rewards for the devices are estimated through observations carried out using reinforced learning that determines the operation state of the device. The reward function is computed using a variant function evaluated using the enduring energy and storage metrics of a device. Unlike the other learning methods, this approach operates in variable communication interval retaining the reward from the previous history of the devices. With a distributed WSN support and recursive knowledge of the sensor devices, REH is intended to improve the energy conservation rate with lesser retransmissions. The curtailed number of retransmissions minimizes delay with more preferable ideal devices in a home management system. The performance of the proposed REH is evaluated through simulations considering the following metrics: end-to-end delay, energy utilization, packets forwarded, expected TTL and number of retransmissions. © 2020, The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature.