This is the comparison written as five specific differences rather than as two definitions, because the definitions are easy and the practical consequences are what a site actually has to plan around. Each of the five is a real operational difference somebody in a maintenance department will feel: what causes work to appear, what has to exist first, who is capable of doing it, which failures each one catches, and where each one wastes effort. Taken together they explain why sites run both.
The trigger, and the prerequisites
Preventive work appears because an interval elapsed. Predictive work appears because a measurement moved. From that follows the second difference: preventive needs only an asset list and an interval, while predictive needs an instrument, a baseline of normal, and somebody who can read it. That is why one can start this week on everything and the other cannot.
Who does it, and what each catches
A preventive visit is ordinary maintenance work by ordinary maintenance staff. Interpreting a vibration spectrum is a specialism, held in-house or bought. And the two catch different things: intervals catch wear that accumulates predictably, while monitoring catches a developing fault between visits, and neither catches a failure mode nobody is looking at.
Where each one wastes effort
Preventive wastes visits on assets that did not need them, which is the price of not knowing. Predictive wastes the whole installation on assets whose failure was never expensive enough to justify it, plus the alerts nobody acted on. Both wastes are real, both are avoidable by allocating per asset, and neither is avoided by adopting one strategy across a site. The waste is also asymmetric in how visible it is. An unnecessary preventive visit shows up as an hour on a timesheet and is easy to point at; an unused monitoring installation shows up as nothing at all, because the capital was spent in a previous year and nobody is counting the alerts that were never acted on.
Questions people ask about difference between preventive and predictive maintenance
Which should a small site choose?
Preventive for essentially everything, plus a hard look at whether one or two assets are expensive enough to justify monitoring. Small sites rarely clear the step cost on more than a handful.
Is condition-based maintenance a third thing?
It is usually used as a synonym for predictive, or as the broader family that predictive analytics sits inside. The distinction matters less than the trigger question, which is the same either way.
Do the two need separate systems?
No, and running them separately is a common mistake. One asset list, one record, two triggers keeps the history in one place, which is where every later question gets answered.