Recently, many neural network methods have been proposed for multilabel classification in the literature. One of these recent methods is the Multi-Layer Extreme Learning Machines (ML-ELMs) in which stack auto encoders are used for tuning their weights. However, ML-ELMs suffer from three primary drawbacks: First, input weights and biases are chosen randomly; second, the pseudoinverse solution for calculating output weights will increase the reconstruction error; third, memory and execution time of transformation matrices are proportional to the number of hidden layers. In this paper, Multi-Layer Kernel Extreme Learning Machine (ML-CK-ELM) that uses a linear combination of base kernels in each layer is proposed for Multi-Label classification. The proposed approach effectively addresses the above-mentioned drawbacks. Furthermore, Multi-Label classification data are inherently characterized by multi-modal aspects due to a variety of labels assigned to each instance. Applying a combination of different kernels is the added advantage of ML-CK-ELM that implicitly assesses the inherent multi-modal aspects of Multi-Label data; each kernel can be effectively used to cover one of the modals better than other kernels. The empirical study indicates that ML-CK-ELM shows competitively better performance than other state-of-the-art methods, and experimental results of multilabel datasets verify the feasibility of ML-CK-ELM.