Robotics has emerged as a major field of research and application over the years, and has also found a place in the curriculum of universities. Robotics as a course is challenging both for the teachers to teach and the students to learn as it involves 3D transformations, algebraic and differential equations, etc., which are difficult to understand. Several robotics learning software have been developed have helped to ease the learning of robotics as a subject. A similar attempt was made in developing RoboAnalyzer, a 3D model based robotics learning software that modelled a serial robot based on its DH-parameters. It could perform forward kinematics and show animation and graph plot as outputs. In this paper, further development of RoboAnalyzer is reported in the form of addition of inverse and forward dynamics analyses of a generic serial manipulator. The important contributions of this paper lie in the development of algorithms using an object oriented modelling approach and the Decoupled Natural Orthogonal Complement (DeNOC)-based recursive formulation. A KUKA KR5 robot was modelled in the proposed software, and the results were verified with those obtained using the Dynamic Simulation module of Autodesk Inventor. RoboAnalyzer can be downloaded for free from http://www.roboanalyzer.com and can be used almost instantly.