Grid computing provides environment that can executes job efficiently by sharing resources in heterogeneous environment. Grid computing environment is evolving into wireless grid and mobile grid environments that consider user mobility and integrate mobile devices into wired grid. Recently, mobile devices are increasing by much in performance, and users also increasing. So researches on utilizing mobile devices have being studied, but most of them don't use mobile devices as resources and don't consider characteristics of mobile devices. Therefore, in this paper, we propose a group-based resource selection algorithm that supports fault-tolerance in mobile grid and considers various characteristics of mobile devices. Our proposed algorithm 1) considers remaining battery power, mobility, and performance of mobile devices, 2) calculates rank of mobile devices and classifies to K groups, and then, 3) assigns a job to top-N groups concurrently for fault-tolerance. Ranking and grouping of mobile devices are performed to consider dynamic characteristics of mobile devices whenever a job is submitted.