This paper includes description of the new approach to the thermal problem solution using modern two-dimensional and three-dimensional neural networks. The mentioned neural networks are used to modelling of , inter alia, the Dual-Phase-Lag (DPL) heat equation [1]. The Dual-Phase-Lag equation is one of the most accurate model that is appropriate for structures which technology node is significantly smaller than about 200 nm [2]. The previous classical approach which uses the Fourier-Kirchhoff differential equation assumes several non-physical behaviours so it could not be use to modelling of thermal dependencies in modern electronic structures [3], [4]. Therefore, this paper presents the solution of the thermal problem obtained using the DPL based heat transfer model. The paper contains also the detailed description of the supervised learning methodology of the artificial neural networks used to solving the mentioned thermal problem and the verification process of proposed algorithm. References [1] Janicki, M.; De Mey, G.; Zubert, M.; Napieralski, A., "Comparison of fourier and non-fourier heat transfer in nanoscale semiconductor structures," in 30th Annual Semiconductor Thermal Measurement and Management Symposium (SEMI-THERM), pp.202-206, 9-13 March 2014 [2] M. Zubert, M. Janicki, T. Raszkowski, A. Samson, P. S. Nowak, K. Pomorski. The Heat Transport in Nanoelectronic Devices and PDEs Translation into Hardware Description Languages. Bulletin de la Societe des Sciences et des Lettres de Łódź. Serie: Recherches sur les Deformations 2014 [3] T. Raszkowski, M. Zubert, M. Janicki, A. Napieralski, "Numerical solution of 1-D DPL heat transfer equation", Proc. of 22nd International Conference Mixed Design of Integrated Circuits and Systems MIXDES 2015, 25-27 June 2015, Torun, Poland, pp. 436-439. [4] M. Janicki et al., Comparison of Green׳s function solutions for different heat conduction models in electronic nanostructures. Microelectron. J. (2015), http://dx.doi.org/10.1016/j.mejo.2015.07.008