You are asking quite a complex question.  Let us break it down. 

The first question to consider is what is a nervous system used for?  

Well an obvious function is to sense the outside world, and respond to it.  In this light, the need for a nervous type function is not restricted animals.

Bacteria for example display a process known as quorum sensing - they send and receive signals and change their behavior accordingly.

Similar types of behavior occur in unicellular eukaryotes.  Do they have a nervous system?

Well not in the traditional sense, but it serves many of the same functions, it coordinates activities and enables cells to adapt (usefully) to their environment - it can lead to cell movement, toward nutrients and away from toxins. 

In a sense, part of your question is about why are there are (so many) different types of organisms - this is relatively simple to answer. 

If there was a single uniform environment, we might  expect a single optimal organismic strategy - a single type of organism. 

But environments are complex and highly non-uniform; there are many solutions to various ecological opportunities – these opportunities are termed ecological niches. 

An active predator needs different types of behaviors compared to a passive feeder or a parasite; these adaptations require different types of coordinating (that is, nervous) systems.  But what a population of organisms can evolve into depends upon what it already is. 

As you probably realize, the overwhelming evidence supports the idea that all living organisms share a single common ancestor.  In the same way, all eukaryotes share a common ancestor that lived about 1.2 to 1.5 billion years ago. 

So now, back to the evolution of nervous systems. 

Even the simplest "colonial" organisms, like sponges (which have no neuronal cells), have ways of coordinating their behavior: and they use molecules that act as neuromodulators in more complex animals [see this reference.]

A key point related to your question is whether the ancestor of all animals, the "eumetazoan" (↑) had a nervous system or not. 

If it did, then the nervous systems of the animals evolved from it would be expected to be derived, via mutation, drift, and selection, from that original nervous system.  This would be considered a “monogenic” model, in which a basic nervous system evolved once, and then was adapted in various ways in various lineages (see this reference.)

An alternative model (see this reference) is that the eumetazoan was more like a sponge, and did not have a distinct nervous system – what it did have was many of the components required to respond to external stimuli. In this case, neurons and nervous systems in different lineages arose independently - an example of convergent, polygenic evolution.

In both cases, the appearance of neurons from cells capable of responding to external stimuli would, in turn lead to networks of cells that could control, at first local, then global, and later increasingly complex) behaviors. 

What makes the two scenarios difficult to distinguish are two facts.

First, all extent organisms are equally ancient, and we need to make assumptions based on data and hypotheses, but difficult to establish unambiguously, as to which most closely resemble the common eumetazoan ancestor.

Genes can be lost, and new genes born during the hundreds of millions of years since various lineages shared a common ancestor (see above).

The second is that the even simplest eukaryotes (for example, sponges), are quite complex cellularly and genetically [reference]. 

They, and the presumptive eumetazoan, already had many genes, encoding proteins, that could be used to build neurons and nervous systems.  

In any case, the nervous system that evolved suited the "needs" of the organism. 

It is worth remembering that that lineages of organisms go extinct and survivors diversifying over time. 

What a population of organisms can diversify into is determined in large measure by its ancestors. 

Organisms that are descended from organisms with complex nervous systems can get more complex or less complex.  

What we assume is that the form their nervous systems took was a function of the components present at the time, the variations that arose, and the selective pressures the various niches available generated.  

Let me know if this answer "fails to satisfy", and exactly what remains confusing.