Wireless sensor network (WSN) is composed of a large-scale self-organizing energy constrained sensor nodes, which is really a network of energy severely constrained in nature. Each sensor node deploying in the detection area has the capability to sense the environment, process the data and relay it towards the sink node or base station through multi-hop wireless communication, and thus complete the deployment area measurements in real time. WSN have very broad application in military surveillance, environmental monitoring, transportation and so on, which have gradually developed into the focus of attention at home and abroad.WSN have been widely applied in many fields while also facing great challenges. How to effectively use nodes energy to extend the lifetime of the entire network should be a critical consideration in network design. NUM (Network Utility Maximization) as an effective tool for system analysis of network architecture, and it is also suitable for the design of various of stable distributed resource control algorithms in WSN.In general, WSN application performance is often evaluated with the amount of data collected by the sink, which could be indirectly measured by the sensor node rate allocation in the network. The larger the rates of the sources are, the better network application performance is and the shorter of the lifetime of WSN, and vice versa. Thus, there is an inherent tradeoff between network application performance and its lifetime. In order to enforce fairness on source rates, we adopt network utility function to measure WSN performance based on the NUM framework. By introducing a system weighing parameter, we have established two convex optimization models that jointly maximizing network utility and network lifetime. The innovation of the proposed model in this paper is that we use the more restrictive MAC constraints instead of link capacity constraints as the general limitations of the feasible region, and also consider the reliable transmission of data based on end-hop retransmission mechanism. Then by the Lagrange dual decomposition method, we have developed the corresponding partially and fully distributed optimization algorithm. Simulation results show that by selecting the appropriate weighing parameter, the proposed models can effectively adjust the data transmission rate of the sensor nodes within the network to achieve a balance between network performance and network lifetime.In recent years, the maximum lifetime routing problem in WSN has received increasing attention in recent years. Another part in this paper, based on differentiation of weighted data fusion mechanism to eliminate redundancy, we have improved the previous model that jointly maximizing network lifetime and minimize the end-to-end packet delay.Lagrange dual decomposition method is adopted to decompose the primal problem into several sub-problems which could be executed on each node, and the corresponding fully distributed algorithm is designed. In numerical simulation part, by the comparison with related research, we find that the algorithm can effectively balance node energy consumption, prolong the lifetime of the network, and reduce end-to-end data packet delay.