In this article, we introduce a kernel-based consensual aggregation method for regression problems. We aim to flexibly combine individual regression estimators r_1 , r_2 , ..., r_M using a weighted average where the weights are defined based on some kernel function. It may be seen as a kernel smoother method implemented on the predictions given by all the individual estimators instead of the original input. This kernel-based configuration asymptotically inherits the consistency property of the basic consistent estimators. Moreover, numerical experiments carried out on several simulated and real datasets confirm that the proposed method mostly outperforms or at least biases towards the best candidate estimator of the combination.