Over the past two decades, more than 160 million Chinese rural workers have migrated to cities to work. They are separated from their familiar rural networks to work in an unfamiliar, and often hostile environment. Many of them thus face significant mental health challenges. This paper is the first to investigate the extent to which migrant social networks in host cities can mitigate these adverse mental health effects. Using a unique longitudinal survey data of Rural-to-Urban Migration in China (RUMiC), we find that network size matters significantly for migrant workers. Our preferred IV estimates suggest that one standard deviation increase in migrant city networks, on average, reduces the measure of mental health problem by 0.47 to 0.66 of a standard deviation. Similar effects are found among less educated, those working longer hours, and those without access to social insurance. The main channel of the network effect is through boosting confidence and reducing anxiety of migrants.