We implement a DualOSCCF (Dual Operating System Collaborative Computing Framework) on ARM PB11MPCore four-core embedded platform, in which Linux and aCoral1 are configured to run. aCoral is a small size realtime operating system (RTOS), which is capable of improving the efficiency of computation-sensitive applications. Linux is mainly responsible for I/O and user interaction, while aCoral is only for computation jobs, such as large-scaled matrix computing. DualOSCCF provides inter-core interrupt mechanism, communication mechanism between Linux and aCoral (Dual-OS), task scheduling and load balancing strategies in multicore environment. When evaluated over matrix multiplication, DualOSCCF is capable of improving the parallelism and throughput for the computation-intensive applications by task decomposition and effective load balancing. Moreover, the collaborative mechanisms provided by DualOSCCF specifications can simplify the application development.