Abstract: Aiming at the problem of low efficiency and slow search in the DGPS ambiguity resolution process, adaptively improve the chicken swarm optimization(CSO) and apply the improved chicken swarm optimization(ICSO) to the fast search of the integer ambiguity, a Kalman filter is used to obtain the float solution and its covariance matrix of the double-difference ambiguity. The Lenstra-Lenstra-Lovasz (LLL) decorrelation algorithm is adopted to decorrelate the float solution and its covariance matrix, thus, the correlation of each ambiguity float estimation can be eliminated. With a fixed baseline length, ICSO is used to search the optimal solution of the integer ambiguity. Simulations are performed using classic examples, the results show that compared with the literature ［5］, the improved chicken swarm optimization can effectively improve the search speed and the success rate of the solution.