Since the 1800s, there have been calls in the literature for the early diagnosis of cerebral palsy (CP). However, diagnosis still often occurs late, from 12 to 24 months in high income countries and as late as 5 years in low resource settings. This is after the optimal timeframe for applying interventions which could harness neuroplastic potential in the developing brain. Multiple barriers exist which affect clinicians’ confidence in diagnosing CP early. These range from the lack of definitive biomarkers to a lack of curative treatments for CP. Further barriers to diagnosis are proposed including; (a) difficulty finding a congruent fit with the definition of CP in an infant, where expected activity limitations might not yet be apparent; and (b) differences in the presentation of motor type and topography classifications between infants and children. These barriers may affect a clinicians’ confidence using “pattern recognition” in the differential diagnosis process. One of the central tenets of this paper is that diagnosis and classification are different, involving different instruments, and are more accurately conducted separately in infants, whereas they are fundamentally interconnected in older children and inform therapeutic decisions. Furthermore, we need to be careful not to delay early diagnosis because of the low reliability of early classification, but instead uncouple these two processes. Ongoing implementation of best practice for early detection requires creative solutions which might include universal screening for CP. Implementation and accompanying knowledge translation studies are underway to decrease the average age of diagnosis in CP.