Handhold Universal Remote Controller (URC) is bulky, confusing, and not always at hand. Wearable controller is compact, intuitive, and always available. Wearable controller share some same issues as handhold controller such as line-ofsight and operation angle of infrared signal, one-for-all encoding method for control command, etc. Meanwhile, different with the handhold controller, the wearable controller with limited computing resources like memory. Therefore, the memory cost on preinstalling the control command for different brands is unaffordable for the resource-constrained wearable device. In this paper, we present a brand-independent low memory footprint method for resource constrained wearable remote controller. A middle device putting in front of each home appliance is used to relay the RF signal to infrared signal. A signal encoding and decoding method are proposed to learn and generate various kinds of infrared control command. In the feasibility experiment, a real test bed is deployed in the home scenario. As a result, different kinds of home appliance can be successfully controlled using a finger-worn controller with an 8 bit MPU and 4 KB RAM. Therefore, the proposed method is promising to facilitate the home automation using wearable controller.