With the development of multi-agent technology, the collaborative localization and mapping has emerged, to enhance the accuracy and robustness in challenging environment. In this study, a sequential graph optimization model of collaborative localization and mapping (CLAM) is proposed, aiming to optimize the navigation state estimations and realize the collaborative localization (co-localization) and collaborative mapping (co-mapping) in a sequential manner, for a multi-UAV (Unmanned Aerial Vehicle) with multi-sensor application. Within the collaboration framework, the state models with the GNSS, INS, UWB and Camera measurements for multi-UAV are constructed and integrated. Co-localization is achieved through the GNSS/INS/UWB fusion in a tightly coupled manner. Co-mapping is obtained by multi-UAV cameras fusion. Dual-UAV experiment is conducted. And the co-localization, co-mapping and sequential CLAM are qualitatively analyzed and quantitatively compared. The results show that compared with only co-localization and only co-mapping, CLAM can improve the accuracy of the mean by \(23\%\) and \(3\%\) , respectively.