Although the SC-Flip decoder can significantly avoids the effects of cumulative errors by flipping the first erroneous bit of the traditional SC decoder, the computational complexity of the SCF increases simultaneously. In this letter, to reduce the computational complexity of the SCF decoder, we investigate the log-likelihood ratio (LLR) values and their distribution related to the erroneous bit-channels, then replace the LLR sorting method of SC-Flip with a new value. By setting a threshold for the bits to flip, we proposed a SC-Flip decoder using effective flip bits (SCF-EF). Simulation results demonstrate that for practical code lengths (1024, 512), SCF-EF can achieve 21.37% complexity reduction at SNR of 0.5dB with negligible loss compared with the BLER performance when correcting all single-errors.