Random Early Recognition (RED) is really a broadly deployed active queue management formula that increases the efficiency from the network when it comes to throughput and delay. The potency of RED formula, however, highly depends upon appropriate setting of their parameters. Furthermore, the performance of RED is very responsive to abrupt changes within the traffic load. Within this paper, we advise a Careful Adaptive Random Early Recognition (CARED) formula that dynamically varies maximum drop probability in line with the degree of traffic load to enhance the general performance from the network. According to extensive simulations conducted using Network Simulator-2 (ns-2), we reveal that CARED formula cuts down on the packet drop rate and achieves high throughput when compared with RED, Adaptive RED and delicate Adaptive RED. Unlike other RED based algorithms, CARED formula doesn’t introduce new parameters to attain performance gain and therefore could be deployed with no additional complexity.