The Computational Brain, Patricia Churchland and Terrence Sejnowski, 1994, ISBN 0-262-53120-8.
This is a good primer if you want to learn computational neuroscience. Explains brain structure along with computational models. It could be described as 'informed connectionism'.
Theoretical Neuroscience: Computational and Mathematical Modeling of Neural Systems, Peter Dayan and L. F. Abbott, 2001, ISBN 0-262-2041995.
Destined to be the definitive textbook. It's a difficult book because the subject is hard. If you have trouble with this book, try the books by Wilson and Rieke et al (see below), then come back to this book.
Spikes, Decisions, and Actions: The Dynamical Foundations of Neuroscience, Hugh Wilson, 1999, ISBN 0198524307.
You need to understand differential equations to do realistic simulations of neurons. If you have three semesters of calculus, but are not yet comfortable with differential equations, this book will be useful. After reading the first half of the book, you can go to the relevant chapters in Dayan and Abbott (above). The last half of the book uses some nonstandard models. You should probably study the standard models in Dayan and Abbott first.
Spikes: Exploring the Neural Code, F. Rieke, D. Warland, R. van Steveninck, W. Bialek, 1999, ISBN 0-262-68108-0.
A book on neural coding. Provides more explanation than the chapters from Dayan and Abbott (above).
Biophysics of Computation: Information Processing in Single Neurons, Christof Koch, 1999, ISBN 0-19-510491-9.
This book is very advanced, yet fundamental. Study Dayan and Abbott (above) first, so that you can really appreciate the book.
Methods in Neuronal Modelling, Christof Koch and Idan Segev, 1998, ISBN 0-262-11231-0.
An an anthology of case studies spanning the levels of structure in neuronal models. Once you start on a project, check this book for appropriate methods. It could prevent some mistakes.