Detecting chaos in time series data is challenging. A new technique was developed to reconstruct chaotic attractors in phase space using multidimensional data. This method was tested on various chaotic systems and showed more accurate results than traditional methods, even for non-chaotic and noisy systems.