The development and testing of object recognition algorithms for range-finding devices requires range-data for a wide variety of scene objects and environments, and further requires an appreciation of the accuracy associated with range-data. A method for generating pseudo (simulated) range-data, based on geometric scene models and modified z-buffer algorithm, is presented in this work. The method provides a versatile and flexible tool for research on 3-D object recognition. Initial investigation of the accuracy associated with range-finding devices is considered. An analytical model for a range-finding device is developed using a homogeneous matrix representation. Differentiation of the resulting object position coordinates is utilized to examine the error in the object position coordinates resulting from relevant device mechanism inaccuracies.