Abstract: Community structure is a common property in many networks. Traditional community detection algorithms aim at discovering all communities in a network and the running time of these algorithms usually depends on the size of the entire network. For big networks，it is difficult to discover their global community structure due to the high time complexity. Local community detection surges as a kind of methods which start with a given node that is already known to be in the target community，and the goal is to uncover the remaining nodes in the community. This kind of algorithms usually starts from a given node and gradually expands outward without requiring global network structure. The running time of these local communities detection algorithms depends on the size of the community rather than that of the entire network. Thus local community detection has important application significance in big network analysis tasks. So far the research of local community detection has drawn a lot of attention and many algorithms have been proposed. Most of the existing local community detection algorithms require users to specify the parameter values manually and have low accuracy. For solving this problem，a new local community detection algorithm is proposed in this paper. Firstly，we propose a new common neighbors based similarity measurement with weighted neighbor nodes，which is employed to calculate the similarity between two nodes in a network. Secondly，based on this similarity measurement，a new local community quality metric is given. We discover a local community of the given node by iteratively adding a node from shell node set to the target community which has the largest embedded degree with the current local community，while ensuring the gain of the current local community quality metric is not negative. Extensive experimental results on both synthetic and real－world network datasets show that the proposed algorithm is highly effective at local community detection compared with Clauset，LWP and GMAC algorithms.