Facial paralysis is a highly burdening condition, resulting in a patient's inability to move his musculature on one or both sides of his face. This condition compromises the patient's communication and facial expressions, and thus dramatically reduces his quality of life. Current treatments for facial paralysis rely on invasive surgical procedures, including nerve transfers or dynamic muscle transfers which both aim for dynamic reanimation of the paralyzed face. Advances in soft robotics offer promising alternatives to restore facial movements without the need for such invasive approaches. This thesis proposes a functional implant to restore facial movements post-paralysis, by placing Dielectric Elastomer Actuators (DEAs) in the paralyzed side of the face and then synchronizing it to the activity of the contralateral healthy side. A significant advantage of this method over traditional therapies is that it eliminates the need for nerve grafts or muscle flaps, thereby avoiding the lengthy waiting periods between surgical procedures, eliminating the need for surgical intervention on other parts of the body and overall reducing the patient's recovery time. Implementing DEAs that mimic natural muscles has been proven difficult, as DEAs provide in-plane expansion when actuated, while natural muscles contract upon stimulation. This research focuses on overcoming the challenge of directional deformation inherent to DEAs. To address this, fiber-reinforced DEAs with uni-axial and structured configurations were developed. Uni-axial fiber-reinforced DEAs demonstrated up to 75% higher strain, a doubled force output, and a 47% improvement in response time compared to non-reinforced DEAs. Furthermore, a comprehensive model is introduced that captures the mechanical interactions between the fibers and the elastomer membrane, thus identifying critical parameters and predicting fiber buckling instabilities. While buckling is generally considered an undesirable instability, this research explored how controlled buckling of fibers can be leveraged to enhance actuation performance. New design possibilities are also explored for artificial muscle implants - notably, this thesis introduces a novel class of contractile silicone-based DEAs reinforced with fibers designed to guide and enhance contraction along a specific axis. Finally, this thesis presented a novel, less invasive approach for facial movement restoration using DEAs as artificial muscles. The primary contribution lies in the development of a dynamic prosthesis that synchronizes artificial muscle movements on the paralyzed side of the face with the activity of the healthy side through a neural interface. By integrating non-invasive EMG electrodes for real-time control, the prosthesis replicates natural facial expressions. Additionally, the thesis explores the design and implementation of DEAs for restoring blinking, demonstrating their potential as a functional alternative to invasive procedures for patients with facial paralysis. This work highlights the potential of DEAs to bridge the gap between soft robotics and biomedical engineering, providing applications for medical prosthetics and beyond.