Electroactive polymer (EAP) actuators, also known as artificial muscles, have remarkable properties such as low energy consumption, low weight, low actuation foot-print, compliance and bio-compatibility. Several methodologies have been proposed to model and analyse their quasi-static bending behaviour with negligible attention paid to their dynamic behaviour. We, therefore, report on an enhanced methodology to model their highly non-linear bending behaviour by treating them as smart and soft robotic manipulators. The methodology consists of an inverse kinematic model and a dynamic model. The proposed methodology accurately estimates the EAP actuator's whole shape deflection using optimization-based inverse kinematic solutions integrated with an electro-mechanical dynamic model. The experimental and numerical results are presented to show the effectiveness of the soft robotic manipulator model in estimating the highly non-linear bending behaviour of the polypyrrole electroactive polymer (PPy-EAP) actuators. The proposed methodology can easily be extended to other bending type actuators and active smart manipulators. © 2014 Elsevier Ltd. All rights reserved.