Based on complex networks, the innovation strategies of enterprises or governments are analyzed by using asymmetric evolutionary games. The evolutionary game model is considered to be a better way to promote collaborative innovation between large and small enterprises. First, a model is established based on the replicator dynamic equations. Then, based on complex network and computer simulation technology, a network evolutionary game model with improved strategy update rules including environmental preference utility is designed. The results show that the conclusions of the mathematical and network models of evolutionary stable strategies (ESS) are the same. The complex network model can provide more detailed information on the evolutionary processes of enterprises, and the parameter values can be adjusted to analyze the evolution sensitivity. In a uniform or non-uniform environment, the effect of environmental preference on evolution is weak, indicating that the effect of collaborative innovation on asymmetric enterprises is weak. The initial probability of enterprise collaborative innovation is the key to ESS. A dynamic model of asymmetric replication factor and an evolutionary game model composed of two participants of large technology enterprises and SMEs are established. Complex networks and environmental preferences are involved in evolutionary games to better analyze the ESS.