Recently there has been renewed interest in the application of photoplethysmography signals for cardiovascular disease assessment. Photoplethysmography signals are acquired non-invasively using visible and infrared light passed through the finger pulp. Unfortunately, this method commonly suffers from many forms of interference and distortion such as; baseline wander, mains-line interference and random spikes or other such artifacts. This paper presents a new approach for effective filtering of the photoplethysmography signal. Specifically, a cascaded filtering method for removing the artifacts from photoplethysmography signals based on the median and polynomial filters (MdPF) is proposed. Recordings from the PhysioNet database are used to validate the proposed method. Our experimental results show that the performance of MdPF cascaded filtering method is more effective than other current methods alone in removing artifacts from photoplethysmography signals. Root mean square error measurements are used for comparison purposes. This paper follows from previous work on median based method for baseline wander removal in photoplethysmogram signals.