Dear Vincent Lecoeuche,Thank you for the suggestion.Its a great help for me.I will search the details!!

Possibly with a piezo actuator

Alternative Method is to put the Fabry-Perot in an isolated box and heat it above room temperature (~30°Deg) via a closed loop temperature control system.

This setup is simple, rugid and cost-effective.

Dear J.H.R. Lambers ,Thank you for the suggestion.

Dear Bernd Markus Bertschinger, seems easy compare to other methods, could you please give me some references ? ,thanks 

Dear Naveed,

please regard

"Thermal tuning of optical cavities for parametric instability control "

for further information.

The paper deals with thermal tuning with regards to a mirror. But it highlights the advantageous with regards to a cooling cycle. 

With regards to optical-cavities, please consider:

"Tunable Grating Fabry-Perot-Based Wavelength Add/Drop Multiplexer in Silicon Photonics "

Cheers

http://nanophotonics.eecs.berkeley.edu/Publications/Conference/files/4408/Yeh%20and%20Wu%20-%202013%20-%20Tunable%20grating%20Fabry-Perot-based%20wavelength%20addd.pdf

http://www.ligo-wa.caltech.edu/~richard.savage/Papers/techpapers_degallaix.pdf

Dear Bernd Markus Bertschinger,Thank you for detailed references.Its a great help for me !

Alternately you could pass light from another reliably stabilised laser through the filter and feedback  (PZT or temp) to stabilise the cavity. You would need to separate the different wavelengths with filters.

Thank you dear Richard, I will try this method. Could you please give me some references ?

It would help if you could tell us what level of stability you need - both in terms of optical frequency and the corresponding temperature changes in the Fabry-Perot.  Do you need stability for a few seconds or over several days?  Frequency stability of a few Hz, kHz or MHz ?

Bernd's suggestion of  an insulated box and heater is worth pursuing.  It is easy to implement, low cost, and should give an immediate improvement.  If it is not sufficient alone it could still be beneficial in combination with other methods.  A box lined with 10 cm thick slabs of expanded polystyrene can work well.  If short term stability is paramount, mount the Fabry-Perot cavity on a metal heat sink to increase the thermal mass, and shield the assembly from convection currents.

If you don't have space for a large insulated box, follow Vincent's suggestion of a Peltier on a heatsink, with insulation and shielding from air currents around the cavity.

Are you able to make or procure a Fabry-Perot  filter with a lower temperature coefficient?  UK National Physical Laboratory use a cavity manufactured from low expansion glass to improve the short-term stability of their optical clocks.  NPL also stress the importance of controlling temperature and minimising vibration for best performance - though they may have more stringent requirements than your application:  http://www.npl.co.uk/science-technology/time-frequency/research/optical-frequency-standards/optical-local-oscillators

Some commercial offerings and background information here: http://www.stablelasers.com/resources.html

To follow Richard's suggestion you need a stable reference laser.  One option is to use a 1542 nm DFB laser locked to the Lamb dip of the P16 transition of acetylene 13C2H2.  https://www.osapublishing.org/oe/abstract.cfm?uri=oe-13-23-9196

There are many other transitions that can be used, including other absorption lines in acetylene, but the P16 transition is a recognized secondary wavelength standard.  Narrow linewidth DFBs are available in this band with output powers up to 100 mW http://www.em4inc.com/node/25

http://www1.bipm.org/utils/en/pdf/CIPM_list_of_approved_radiations.pdf

A fabry-perot filter is not sufficient for laser stabilisation - it is used for frequency narrowing and frequency selection of the transmitted beam. Is that all you care about?

As Alan indicated, more information is needed -  what is the experiment you are putting together? What is the rest of your setup? What kind of laser? Do you have any tuneable elements? Do you have a stable reference to use such as a cavity, atomic reference or another stabilised laser?

If you are only really interested in stabilising the frequency dependence of your filter, the simplest method is one described above where you isolate your optical elements from the environmental temperature fluctuations by putting the laser and filter in a hermetically sealed box. Heating this box to a constant temperature using a temperature PID loop ensures that you are not sensitive to long-term thermal drifts over several hours/days.

If you care also about your laser frequency stability, you will need to adopt a laser feedback setup such as using a stable reference laser or cavity or atomic/similar reference.

Even if your filter is very robust to temperature variations and so has well-defined and stable transmission, if your laser is drifting around you will likely have intensity variations of the transmitted beam. Depending on the scale of your laser frequency drift (e.g. large) and the width of your filter transmission (e.g. narrow), you may find your laser drifts far away from the filter transmission window.