In this paper to achieve higher data rates with high spectral efficiency and high accuracy we designed training sequence sparse channel estimation based on BAT, Cuckoo and Firefly algorithms. By using the above techniques we design a Training sequence channel estimation to reduce the bit error rate, mean square error and accurate recovery of data. The firefly optimization is the promising technique to reduce the bit error rate and to increase the signal to noise ratio to achieve high spectral efficiency Gbps.