In this study, kinematic analysis of a planar slider-crank mechanism having revolute joints with clearances was presented. Joint clearance was modelled as a massless virtual link, and Multi-Layered Neural Network (MLNN) structure was used for approximating the motion of this link with respect to the position of input link. Training and testing data sets for the neural network were obtained from mechanism simulation using the ADAMS software. A genetic algorithm was also used to optimize the design parameters for minimizing the deviations due to clearances. When two joint clearances at crank-pin and piston-pin centers were considered, the effects of these clearances on the kinematic characteristics and transmission quality of the mechanism were investigated using continuous contact model between the journal and bearing at a joint.