In order to solve the problem of that the standard salp swarm algorithm(SSA) has slow convergence velocity and low result precision in the evolutionary process, an improved algorithm, called crazy and adaptive salp swarm algorithm(CASSA), is proposed in this paper. The Tent chaotic sequence is used to initiate the individuals’ position, which can strengthen the diversity of initiate the individuals. The crazy operator is introduced at the food source position to increase the diversity of the population. The adaptive inertial weight is introduced into the follower position update formula to balance the global search and local search ability of the algorithm. The efficiency of the CASSA is evaluated by using statistical analysis, convergence rate analysis, Wilcoxon's test, standard deviations on classical benchmark functions and modern CEC 2014 functions. The results show that the CASSA has better global search ability and solving robustness, and meanwhile, the optimization accuracy and convergence speed are also more powerful than the standard algorithm. Especially, in solving the high-dimension and multimodal function optimization problem, the improved algorithm has better performance.