Recently, the interest of the telecommunication operators and Internet service providers (ISPs) in energy efficiency (EE) for wireless networks has remarkably increased. The boost in energy prices, the continuous increase of the user numbers, the pervasion of wideband access, and the diffusion of offered services have motivated this interest. The challenge is to minimize the transmitted energy without compromising the quality of service (QoS) and the network performance. In this paper, a new efficient Orthogonal Frequency Division Multiple Access (OFDMA) resource allocation for Long Term Evolution (LTE)-advanced systems based on cross-layer optimization between Physical (PHY) and Medium Access Control (MAC) layers is proposed. A new solution based on Rate Adaptive (RA) principle is presented to achieve better EE resource allocation. The purpose behind the new model is to increase the capacity of the system as much as possible while trying to decrease the total transmit power. Hence, the EE which is equal to the capacity to total transmit power ratio is increased. As reducing the transmit power for the transmission of a given capacity reduces the power consumption, the greenhouse gas (GHG) emission will be then reduced. On the other hand, the service differentiation is highly considered to provide high QoS support. The users are differentiated according to different service classes and different data rates. The proposed cross-layer-based solution has then to allocate the system resources in an optimal way while satisfying the users having high QoS requirements.