Abstract: In order to improve the success rate and search efficiency, the paper suggests applying modified TIKHONOV regularization method to re duce ill*.condition of normal matrix which is formed by singl e epoch observations to get more reliable and nearer real solution. Then the wor k load for searching integer ambiguity is reduced which is composed of the float ambiguity and covariance. To use genetic algorithm can search the best candidat e integer set depending on a fitness function. Some results show that the method could improve the success rate and efficiency for determining single epoch carr ier phase. And the method provides some new ways for applying single epoch measu rements to fulfill high precision dynamic positioning. 