Flocking behavior is widely used in virtual reality, computer games, unmanned vehicle, robotics and artificial life. However, coordination of multiple flocking behaviors to accomplish such tasks remains a challenging problem. This paper reports some progress for implicit coordination and gets swarm intelligence as works based on the flocking behavior. It consists of two parts. In the first part, we study on the pattern formation problem with avoiding complex constraints, that is how can a group of agents be controlled to get into and maintain a formation. The second part considers the studies that use adaptation strategies in controlling multiple agents based on probabilistic methods. Specifically we investigated (1) how probabilistic method is used to reorganize generate group (flocking) behaviors, and (2) how adaptation at the individual level is used to make multiple agents respond to obstacles in the environment.