A novel technology for rapid quantification of heavy metals in water using an electronic voltammetric tongue and deep neural networks is proposed. The electronic tongue has 4 working electrodes (gold, platinum, antimony and glassy carbon) and works under the principle of anodic stripping voltammetry (ASV), the device has an autonomous cleaning system, oriented to automate the process of heavy metal characterization in real industrial processes. A total of 172 different artificial samples were generated with atomic absorption spectrophotometry standard solutions of Pb(II), Cd(II), Hg(II), As(II) and Cr(III), the samples contained 1 to 5 of the metals in different concentrations between 1 and 20 ppm. Finally, the samples were analyzed with the electronic tongue. The obtained data was randomly divided into training and test sets for the training of deep learning hybrid models composed mainly of 1D convolutional layers and fully connected layers. After optimizing the neural network architecture, independent models were obtained for each of the metals with mean absolute errors of 1.523 ppm, 0.767 ppm, 0.984 ppm, 1.308 ppm and 1.370 ppm in the test data sets for As(II), Cd(II), Cr(III), Pb(II) and Hg(II) respectively. The electronic tongue showed good predictive power for the quantification of the analytes of interest