Differential Evolution (DE) is a simple but powerful population-based stochastic optimization algorithm. Owing to its simplicity, easy implementation and excellent performance, DE has been wildly applied in scientific and engineering areas. However, there are still some inconveniences and weaknesses in DE algorithm, such as the inconveniences in the choice of proper control parameters and the defects existing in a given mutation strategy. In this paper, a new DE variant, called Parameters with Adaptive Learning Mechanism Differential Evolution (PALM-DE), is proposed to tackle the inconvenience in control parameter selection as well as to enhance a former mutation strategy. The new variant is verified on 44 commonly used real-parameter single objective benchmark functions selected from CEC2013 and CEC2014 competitions. Several recently proposed well-known DE variants are also contrasted in the paper, and the experiment results show that the proposed PALM-DE algorithm is competitive in comparison with these DE variants. An attempt to enhance the performance of PALM-DE by employing linear population size reduction is also presented, and the performance is still competitive.