<p>Rolling bearing is the key component of rotating mechanical equipment, whose performance will affect the health and service life of the equipment significantly and directly. Therefore, the fault diagnosis of rolling bearing failures is necessary for avoiding the accidents caused by them. Because of their fault features being masked by noise, it is difficult to acquire the features and distinguish the types of faults by the nonlinear vibration signals of rolling bearing directly. To solve this problem, a fault diagnosis method for rolling bearings based on parameter optimized variational mode decomposition (VMD) and one-dimensional convolutional neural network (1DCNN) was proposed. Firstly, the whale optimization algorithm (WOA) improved by an adaptive convergence parameter and perturbation factors were applied to optimize the parameters of VMD. Secondly, the parameter optimized VMD was applied to obtain the required intrinsic mode functions (IMFs) from the vibration signals of rolling bearings to denoise them. Finally, densely connected 1DCNN model adding with the proposed position-encoded temporal attention module (PTAM), named PTAM-Dense-1DCNN, was used as the feature extraction and fault classification network for fault diagnosis. The result of the experiment performed on two experimental bearing datasets have shown the effectiveness of proposed methods, whose superiority over other representative methods has also been demonstrated by comparative experiments in this research.</p>