A two-stage optimisation method is proposed to dynamically balance a planar slider-crank mechanism to minimise the shaking force and shaking moment and synthesise its link shapes. In the first stage, the dynamic balancing is achieved by optimally distributing masses of crank and connecting rod using the equimomental system of point-masses. Their shapes are synthesised systematically by closed parametric curve, i.e., cubic B-spline curve corresponding to the balanced mechanism's inertial parameters in the second stage. When such optimised mechanism is used in a multi-cylinder inline engine, the loads on main bearings get reduced. The optimisation problem thus formulated to minimise shaking force, shaking moment and bearing loads as well as to find optimum link shapes are solved using the popular evolutionary optimisation algorithm, i.e., genetic algorithm.