Sit-to-stand (STS) transfer is an easy everyday activity for healthy people, but it is a challenging task for the elderly, and patients with lower-limb weaknesses. This article presents an intelligent assistive walker, which assists the user to accomplish safe STS transfers. An STS intention recognition method is proposed based on an optimized adaptive network-based fuzzy inference system (ANFIS), where the walker is able to accurately detect a voluntary STS and to apply an appropriate control accordingly. The walker is controlled by using a proxy-based controller with respect to the detected human intention, that is measured using thin-film force sensing resistors. During the STS transfer, the assistance force is generated on-line using a virtual spring-damper between the proxy and the linear actuator of the assistive walker. The stability of the proposed controller is theoretically analyzed, and verified by simulations. Experiments with five young healthy subjects and one elderly subject demonstrated that the proposed assistive walker is able to efficiently follow the user’s intention to complete stable and compliant STS transfers.