Today, many replacement algorithms have been proposed to improve the performance of Web caching. Most suggested algorithms perform well, but they require storing many parameters and need a long time to select the document for replacement. These can be a problem, because ultimate object of replacement algorithms is to make simple and out-performed approach and to make this algorithm be implemented in current proxy cache easily. In this paper, we propose a new algorithm, called LRU based Small Latency First Replacement (LRU-SLFR), which combines LRU policy with real latency to achieve the best overall performance. We employ the simplicity of LRU and compensate the LRU algorithm with real latency parameter and access-count. In our scheme, the proxy remembers the latency and the access-count of documents. Based on this information, we make the linked-list as the LRU policy does and make a group of documents that have the same access-count from the tail of the linked-list. We call this group; Same Conditional Group Window (SCGW). If proxy must replace a certain document, proxy that uses our algorithm replaces the document that takes the smallest time to load in the SCGW. The computations associated with the replacement policy require only a small time. Using several simulations, we show that LRU-SLFR out-performs the LRU policy in latency reduction, hit ratio and network traffic, without serious calculation of several parameters.