Abstract The enormous traffic of machine-type communications (MTC) expected over 5G exacerbates the limitations of access schemes currently under investigation in the literature. This scenario becomes more challenging when considering smart city environments, which introduce further issues due to the heterogeneity in the level of residual battery energy of involved machines. Novel solutions are, therefore, required, which aim at drastically reducing the collision probability of devices with critical level of residual battery energy. In this paper, we propose a virtual code resource allocation (VCRA) approach which extends the code-expanded strategy to support a high number of devices simultaneously accessing the system. Besides, a virtual resource allocation scheme to guarantee energy-priority in the access procedure is introduced. The idea behind our proposal is the definition of different access levels that exploit disjoint sets of access codewords, properly tailored to guarantee high capacity for each access level. Simulation results show to the effectiveness of our scheme in terms of ( i ) reducing the collision probability of machines with limited battery capabilities also in scenarios with very high cell load and ( ii ) enhancing the efficiency with respect to legacy code-expanded strategy.