In multitasking and time sharing operating system the performance of the CPU depends on waiting time, response time, turnaround time, context switches from the process mainly depends on the scheduling algorithm. Shortest Job First(SJF) is one of the most widely used scheduling algorithm because it give less Average waiting Time compare to other Scheduling algorithm but this algorithm have some disadvantages. Here burst time of processes have to predict (i.e. how much time process will execute in CPU) before CPU start execution. But it is quite impossible to predict the Burst time of process before process complete execution and if processes did not finish their execution in predicted burst time then in which algorithm process will execute? If they execute in FCFS manner Average waiting time may increase and there is a possibility to occur starvation. In this paper based on the experiments and calculations a new algorithm is introduced. In this algorithm the main idea is if processes remain in ready queue after completion of their predicted Burst Time, a new quantum will generate dynamically and each non completed process will get that amount of time quantum to execute in CPU if any process does not completed after execution of that amount of time (Time Quantum) it will go to tail of the ready queue and waiting for it’s turn.