Blackhole attack is considered to be a virulent attack in mobile ad hoc networks (MANETs) that operate on reactive routing protocols. This paper presents an efficient cross layer approach to detect the malicious node that launches the blackhole attack. The proposed scheme uses Adaptive Neuro- Fuzzy Inference System (ANFIS), which combines the benefits of neural networks and fuzzy logic, in the detection process. Experimental results prove that the cross layer approach using ANFIS outperforms the scheme using single layer approach. The proposed scheme produces 100% true detection when Fuzzy C-means (FCM) clustering is used and 95.9% for subtractive clustering. The detection time has also been comparatively reduced. © 2006-2017 Asian Research Publishing Network (ARPN).