Decision is a process which is dependent on environment, knowledge which includes cognitive domain and technology. Observe-Orient-Decide-Act (OODA) for decision has been first given by Fighter Pilot John “Boyd” of US in 1942. In Boyd OODA loop the “Time” of the loop is very important, which means that the OODA loop time should be faster than the opponent OODA loop time to win the conflict/game. Other domains of the decision theory includes game theory, fuzziness, chaos theory, cognitive science, computational data analytics and big data with Artificial Intelligence (AI), etc. The improvised OODA loop also has feedback and feed-forward loops. There are various variants improvised of Boyd's OODA like Team- OODA (T- OODA), Cognitive-OODA (C-OODA), etc. This paper is proposing Fractal-OODA concept which means that each phase like Observe, Orient, Decide and Act has its own internal OODA loop with different timings. At the end as time progresses it will become Fractal-OODA. The Fractal Dimension (FD) of OODA will give the complexity of decision, low FD of OODA (easy or simple decision) and high FD (complex decision). Modern decisions are enabled with AI technologies with Human out-of-the-loop so that decisions are scientific based on AI and probability of errors are less.